ISSUE SPECIFIC
TECHNICAL GUIDE
September 2009
Appendix
1: Modeling/Release Standards
Appendix
2: Copy of Invitations to Panelists
Mediamark Research has been providing
average issue audience information for magazines since 1979. These data, along
with more extensive data on other media, demographic, psychographic and
consumption data, are published every six months.
In mid-2006, MRI began fielding a
companion study to the National Study that moves beyond the average issue measure
to provide audience information for each issue of a title. Based on the
assumption that not every issue of a magazine yields the same audience size,
this Issue Specific Study has as its primary objective the measurement of
audience variability from issue-to-issue. Additionally, the Issue Specific
Study is updated and released weekly to provide users with more up-to-date
information on actual issues of a title.
It is important to note, the Issue Specific Study is not intended to
replace the MRI National Study as the audience currency. The Issue Specific
Study is a companion to the MRI National Study and, in fact, works with this
well-established study.
I.
Methodology
a. Sample
The original launch of this study
included use of Survey Sampling International’s (SSI) Survey Spot Panel. In
January 2008, in an interest to increase the weekly number of completes and the
stability of some of the data, MRI began using also the e-Rewards panel for
sample. Since January 2008, MRI has collected 2,500 completed surveys from both
panel sources (5,000 completed surveys) on a weekly basis. The data from both
samples is merged before release. Below, information is provided for both panel
sources.
As of November 2009, the SSI panel had a total of approximately 1.2 million members and, as of September 2009, the e-Rewards panel had a total of nearly 2 million members.[1]
SSI and e-Rewards utilize different
techniques to recruit panelists. That
said, both panel companies create a process that aims to limit ineffective
survey response/participation (i.e. fraud, etc.). While SSI recruits panelists
by posting ads on thousands of websites, e-Rewards focuses its efforts on creating
partnerships with different companies (i.e. airlines, home stores, apparel
retailers, etc.) and inviting users of those companies to join their panel.
Both companies attempt to broaden the websites/companies that they use to
recruit panel members.
Both SSI and e-Rewards have written
documents that outline membership restrictions and panel management techniques
(SSI’s document is called “Summary of Panel Management Practices” and e-Rewards
has a document called “e-Rewards Answers ESOMAR’s 26 Questions to Help Research
Buyers” ). These documents assert that panelists must:
· Have
clearly and actively indicated their intention to join
· Be
18 years of age or older
· Have
received a welcome message with the opportunity to opt out
· Not
be a duplicate of another panelist (based on email and postal address matching)
· Possess
confirmed validated data (geography and demographics) matched to postal address
files and correct geographic assignment
· Be
assigned a unique panelist ID which is their identifier and can be used for
deduping, recontacts and for post-survey analysis, if needed
· Understand
that survey participation through the respective panel will not serve as a
means to supplement their income
· Agree
to keep confidential information such as new product concepts disclosed to them
in surveys
SSI varies the rewards it issues for
participation based on population segments. For example, younger members
(18-24) receive small cash incentives for participation, while other segments
receive no cash incentive but rather points or charitable donations. e-Rewards’ panelists earn currency or
points that are redeemable for a variety
of rewards.
b. Demographic
Breakdown: Survey Spot and e-Rewards Panels Versus the National Profile
Comparing the demographic breakdown of
the Survey Spot and e-Rewards’ panels against the national profile (using data
from the MRI Spring 2009 National Study) shows the limitations and biases
associated with an Internet panel. For example, while the U.S. adult population
(18+) is split 51.7% female and 48.3% male the corresponding sex split of the
SSI panel shows a greater skew toward females (69.32% female and 30.71% male),
as does the e-Rewards panel (62.35% female and 37.65% male).
The
following tables compare the U.S. demographic profile to that of the Survey
Spot panel (as of November 2009) and the e-Rewards panel (as of September
2009).
|
Age Break |
U.S. Profile |
Survey Spot Profile |
e-Rewards Profile |
|
18-24 |
12.69 |
25.54 |
4.03 |
|
25-34 |
17.88 |
23.85 |
26.31 |
|
35-44 |
18.98 |
18.78 |
25.19 |
|
45-54 |
19.53 |
17.14 |
22.73 |
|
55-64 |
14.63 |
10.09 |
15.27 |
|
65+ |
16.29 |
4.30 |
6.12 |
The Survey Spot panel is skewed younger than the general U.S. adult population. The e-Rewards panel shows shortfalls in the youngest and oldest age cohorts. It should be noted that a small percentage of panelists do not report their age, so there is a small amount of missing information.
|
Employment Status |
U.S. Profile |
Survey Spot Profile (%) |
e-Rewards Profile |
|
Employed |
63.04 |
54.23 |
67.13 |
|
Not Employed |
36.96 |
45.79 |
2.73 |
In the case of the SSI panel, while the proportion of individuals who are not employed is quite close, there are slightly more individuals who are employed part time on the panel compared to the U.S. adult population. Conversely, there are fewer individuals who are employed full time on the panel compared to the U.S. adult population. Again, a small percentage of panelists have not reported their employment status. When looking at the e-Rewards panel, one can see that there is a very small pool of panelists who are not employed. In the case of the e-Rewards panel, there is significantly more missing data on employment.
|
Education Level |
U.S. Profile |
Survey Spot Profile (%) |
e-Rewards Profile |
|
Did Not Grad HS |
14.73 |
9.56 |
0.65 |
|
HS Grad |
31.18 |
28.60 |
8.26 |
|
Some College |
27.54 |
35.06 |
25.26 |
|
Grad College Plus |
26.55 |
25.73 |
63.62 |
The SSI panel around the demographic
education looks quite similar to the general adult population. However, the
e-Rewards panel is clearly skewed in the direction of more education versus the
overall U.S. adult population.
|
Household Income Range |
U.S. Profile MRI
S09 |
Survey Spot Profile (%) |
e-Rewards Profile |
e-Rewards Profile |
|
Less than $20,000 |
13.77 |
28.97 |
Less than $25,000 |
6.06 |
|
$20,000-$29,999 |
9.73 |
13.63 |
|
|
|
$30,000-$39,999 |
9.82 |
11.31 |
$25,000 to $49,999 |
17.64 |
|
$40,000-$49,999 |
9.15 |
8.7 |
|
|
|
$50,000-$59,999 |
8.46 |
7.21 |
$50,000 to $74,999 |
20.24 |
|
$60,000-$74,999 |
10.99 |
7.11 |
|
|
|
$75,000-$149,999 |
27.85 |
9.77 |
$75,000 to $149,999 |
33.83 |
|
$150,000 or more |
10.22 |
1.70 |
$150,000 or more |
10.85 |
Comparing the available information
with the MRI National Study data shows fewer individuals in the higher
household income ranges than in the general adult population in the case of the
SSI panel. Conversely, the e-Rewards panel is skewed more to the upper income
groups.
The disparities between the general
population and the demographic make-up of the Survey Spot and e-Rewards panels
are undeniable. However, these disparities are consistent from week to week of
the survey. It is important to remember that Internet panels, in general, are
non- representative samples of the general population.
c. Sample
Selection Process
Every week, both SSI and e-Rewards,
respectively, select a random national sample of panelists. Both panel companies
pull sample in the same proportions by the four groupings listed below:
4.8% of invitations to males ages 18-24
3.8% of invitations to females ages
18-24
52.1% of invitations to males 25 or
older
39.3%
of invitations to females 25 or older
These panelists are screened against
completed respondents (those who have completed an Issue Specific survey within
the last six months).
Every Wednesday at approximately 3 PM
EST, email invitations to the sample they have pulled for the week for the
Issue Specific Study..
Every Thursday, depending on the number
of completes generated overnight, SSI either sends a reminder to the sample
already invited and/or selects additional, supplemental sample. Supplemental
sample is usually pulled for males only. The number of reminders and/or the
size of the supplemental sample vary depending on the response rate from the
first night (Wednesday) of data collection.
e-Rewards’ initial mailing is frequently sufficient to reach quote, but
occasionally they do deploy further batches of approximately 1,000 invitations
as needed on Thursdays.
d. Invitation
The MRI Issue Specific Study was first
fielded on May 31, 2006. Data from this first field period through the May 24, 2007
field period comprised the first release of this study (this amounts to 52
weeks of data collection).
As detailed above, members of the SSI
Survey Spot and e-Rewards panels are invited to participate in the Issue
Specific Study. Invitations are sent out via email every Wednesday at 3 PM EST.
A copy of each of the invitations is included in Appendix 2.[2] Embedded within the
invitation email is a link to the online survey. Depending on the rate at which
the survey is yielding completes, a reminder email for SSI or e-Rewards may be
sent out sometime on Thursday. This reminder email is identical to the original
email invitation with only a change in the subject heading of the email from
“Media Behavior Study” to “Reminder: Media Behavior Study”.
The survey is open from Wednesday
afternoon each week until Thursday night. Around 5,000 completed surveys are
obtained every week. The final number of invitations that are mailed out each
week to generate that quota of completes varies. Evaluating all of the field
weeks to date produces a cooperation rate range of between 1.21% and 4.32%. The
average range of cooperation for the period that corresponded with the Spring
2008 release is 2.66%.
Each week is an independent sample of
completed surveys. Once a respondent has participated in the survey, he/she is
not eligible to be invited back for six months.
e. Questionnaire
The survey begins with the six-month
screen question using black and white logos for each magazine as the recall
stimulus. This procedure mimics the MRI National Study personal interview.
Approximately 20 logos are placed on the grid on each screen. Logos are placed
randomly within the grid each time the survey is started by a respondent. This
ensures that every logo has an equal chance of being placed in every position
of the survey. Respondents are asked to select which, if any, of the magazines,
whose logos are on the screen, they have read or looked into in the last 6
months. There is a “none of the above” option given on each screen. A
respondent must select at least the “none of the above” option on each screen
in order to advance to the next screen.
Currently, the logos from the Wave 61
National Study field period are being used in the Issue Specific Study. All
logos included in the National Study are included in the Issue Specific Study,
excluding logos for the Spanish language titles. There are a total of 245 logos[3]
included in the study currently. The airline titles were included when we
updated the survey with the Wave 58 logos in November 2007. The airline titles
are arrayed randomly with the other magazines, which is slightly different from
the procedure employed in the National Study.[4]
This is a screen shot of the first set
of screens in the study (the six month screen question). Note, one cannot see
the twenty logos on the screen shot above. Computers have different screen resolutions
and sizes, so this may differ by computer. One can see the scroll bar on the
far right-hand side of the screen shot that indicates there is additional
information (i.e. additional logos) below. The “none of the above” option is
found at the bottom of the logo grid box.
The Sunday magazines are included in
the Issue Specific Study. In this study, the Sunday magazines are measured
directly. These titles are found on a screen separate from the other magazines
with a modified description. Randomization of the Sunday magazine logos is also
conducted so that each title has an equal chance of appearing first in the grid
or in any other position of the logo grid.
Once respondents have indicated the
titles they have read or looked into in the last six months, the survey
proceeds to the issue specific section (see two example screen shots below).
Respondents are shown the most recent covers for those titles they have
identified as having read or looked into in the last six months. For weekly and
bi-weekly titles, the respondent is shown the eight most recent covers. For
tri-weeklies, monthlies and bi-monthlies, the respondent is shown the four most
recent covers.
Covers are ordered from the least
recent (left side of the screen) to the most recent cover (right side of the
screen). Each cover is accompanied by its on-sale date. Additionally, the size
of the covers has been maximized in order to give respondents the most
information about the issue of the title. Therefore, on most covers, the
respondent is able to read the major articles that are represented by headings
on the cover.
There are magazines that produce split
covers for all or particular issues. There are different types of splits:
newsstand/subscriber, geographical, demographic, etc. MRI can only include one
cover per issue in the study. Take the example of a newsstand/subscriber split
in cover. If there is such a split in an issue(s) of a title, MRI requests that
the magazine provide the cover representing the bigger proportion of its
circulation (e.g., if the title has a bigger proportion of subscribers, MRI
requests the subscriber cover in this case). In the Fall of 2009, MRI entered
into a development phase to review the current approach to split covers and is
working on a new standard of rules for such cases.
In the case of geographical splits, MRI
successfully tested an approach for including geographically split covers and
found an effective method to measure such splits. This approach was first implemented in August
2009. For more information, please
contact MRI.
There are many different permutations
of splits/cover treatments employed by magazine publishers. It is important
that MRI shows the covers representing the largest percentage of the magazine’s
circulation. In cases where this is impossible, there are implications for
release of that given issue of the title (see release standards in Appendix 1
and list of non-released magazines and issues in Appendix 4).
In the issue specific section of the
survey, the respondent is asked to select which, if any, of the issues shown on
a screen they have ever read or looked into. Again, a “none of the above”
option is provided to the respondent. And, the respondent must log in some data
on each screen in order to advance to the next screen.
The screen shot above is an example of
a screen for a weekly title. As established above, respondents see the eight
most recent issue covers for weekly and bi-weekly titles. Every week, unless
there is no new issue being published, the oldest cover drops out of the study
and the new cover is introduced sending all the remaining covers down one
position in the line.
This second screen shot example
is of a monthly title. Respondents see the four most recent issue covers for monthly,
tri-weekly and bi-monthly titles. Similar to the example provided for the
weeklies, the monthly covers are ordered from the oldest of the covers (far
left-hand side of the screen) to the newest (far right-hand side of the
screen). In the case of a monthly, approximately every four weeks (or upon the
introduction of a new issue of the title) the oldest cover drops out of the
study and a new one is introduced. The issue specific section of the survey is
followed by two qualitative questions: the “where read” and “how obtained”
questions. These questions are asked generically of a title rather than of each
specific issue. These questions are asked generically so as not to create a
burden for the respondent, which could potentially cause the respondent to
terminate the survey midway through, thereby invalidating the survey.
Only completed surveys are included in
the 5,000 quota each week.
The respondent may select both of the
options if he/she so chooses. In the case where the respondent does select both
options for this question, he/she is asked a follow-up question about where
he/she reads or looks into the magazine most often. The respondent can only
select one of the response options in the follow-up question.
Similar to the description of the
“where read” question above, a respondent may select multiple response options
for the “how obtained” question. In cases where multiple responses are
selected, the respondent is asked a follow-up question to determine where
he/she most often obtains the copy of the title.
The respondent is then asked two
Internet related questions: how often he/she uses the Internet in a week and
where he/she is taking the survey. Finally, the survey ends with a series of
demographic questions:
·
Sex,
·
Age[5],
·
Marital status,
·
Employment status,
·
Education,
·
Household income[6],
·
Race[7], and
·
Hispanic.
The survey takes on average less than
10 minutes to complete (this information applies to both sample sources).
a. Weighting
The following variables were used for
the standardization of the Issue Specific data. MRI used the average of the
first twenty weeks of data collection to develop these standardized targets for
each variable. Standardizing the weights provides consistent demographic
profiles for each week’s sample without substantially impacting the effective
sample size of the study. All variables were standardized within sex.
|
Women |
Men |
|||
|
Weighted Total |
53593 |
Weighted Total |
46407 |
|
|
Hispanic origin |
% |
Hispanic origin |
% |
|
|
Yes |
4.513 |
Yes |
5.268 |
|
|
No
|
95.487 |
No |
94.732 |
|
|
Race |
% |
Race |
% |
|
|
White
only |
87.244 |
White only |
85.726 |
|
|
Not
white only |
12.756 |
Not white only |
14.274 |
|
|
Household income |
% |
Household income |
% |
|
|
Under
$25,000 |
13.316 |
Under $25,000 |
10.889 |
|
|
$25,000
- $49,999 |
23.607 |
$25,000 - $49,999 |
21.338 |
|
|
$50,000
- $74,999 |
20.289 |
$50,000 - $74,999 |
20.496 |
|
|
$75,000
- $99,999 |
13.177 |
$75,000 - $99,999 |
15.450 |
|
|
$100,000
or more |
15.016 |
$100,000 or more |
20.979 |
|
|
Prefer
not to say |
14.595 |
Prefer not to say |
10.848 |
|
|
Education |
% |
Education |
% |
|
|
Graduated
college or more |
40.014 |
Graduated college or more |
49.816 |
|
|
Attended
college |
41.078 |
Attended college |
36.292 |
|
|
Graduated
high school or less |
18.908 |
Graduated high school or less |
13.892 |
|
|
Employment |
% |
Employment |
% |
|
|
Employed
|
64.878 |
Employed |
73.468 |
|
|
Not
employed |
35.122 |
Not employed |
26.532 |
|
|
Marital status |
% |
Marital status |
% |
|
|
Single |
19.814 |
Single |
24.215 |
|
|
Married
|
59.213 |
Married |
60.580 |
|
|
Other
|
20.973 |
Other |
15.205 |
|
|
Age |
% |
Age |
% |
|
|
18
- 24 |
6.270 |
18 - 24 |
5.566 |
|
|
25
- 34 |
20.858 |
25 - 34 |
19.777 |
|
|
35
- 44 |
22.996 |
35 - 44 |
20.275 |
|
|
45
- 54 |
25.644 |
45 - 54 |
21.907 |
|
|
55
- 64 |
17.584 |
55 - 64 |
21.157 |
|
|
65
or older |
6.648 |
65 or older |
11.317 |
|
b.
Examples of data spreadsheets
The following data tables depict how
MRI sees the data generated in the Issue Specific Study. The week-to-week index
information shown here is not included in the deliverable; only the final
indices found on the far right of each table (in the gray shaded column) are
shown. The final index is an average index for a given issue across all the
weeks that this issue was in the field.
The first example is for a weekly
magazine.
Table
1
Example
of Data Input for Weekly Magazine
|
|
5 Days Since On Sale |
12 Days Since On Sale |
19 Days Since On Sale |
26 Days Since On Sale |
33 Days Since On Sale |
40 Days Since On Sale |
AVERAGE
|
|
Issue 1 |
133
|
124
|
120
|
88
|
99
|
153
|
120
|
|
Issue 2 |
78
|
84
|
40
|
66
|
80
|
81
|
72
|
|
Issue 3 |
73
|
59
|
54
|
92
|
96
|
88
|
77
|
|
Issue 4 |
72
|
79
|
108
|
87
|
104
|
84
|
89
|
|
Issue 5 |
115
|
94
|
80
|
118
|
110
|
58
|
96
|
|
Issue 6 |
103
|
81
|
95
|
90
|
46
|
97
|
85 |
|
Issue 7 |
191
|
142
|
147
|
96
|
179
|
118
|
140 |
|
Issue 8 |
106
|
127
|
67
|
133
|
70
|
111
|
102 |
|
Issue 9 |
129
|
54
|
126
|
49
|
112
|
103
|
96 |
|
Issue 10
|
56
|
98
|
73
|
90
|
66
|
102
|
81 |
Table 1 shows indices for each of 10
issues of a magazine over 6 weeks.[8] These replicated
comparisons provide a robust database for producing issue specific
information. For example, issue 1
produced an average index of 120. Compared with all other issues for this
title, this issue did 20% better than average. Note that, in a substantial
majority of instances, issues that do well tend to do well across all or close
to all of the weeks in the comparative database. For issue 1, four out of the
six weeks demonstrated much higher indices than average. The same can be said
of Issue 7 which produced a final index of 140. Conversely, the same
consistency is found when issues do worse than average. Issue 3, for example, produced a final index
of 77, and had six out of the six weeks of measurement yielded indices under
100, as well.
The second example, below, is for a
monthly magazine.
Table
2
Example
of Data Input for Monthly Magazine
|
|
8 Days Since On Sale |
15 Days Since On Sale |
22 Days Since On Sale |
29 Days Since On Sale |
36 Days Since On Sale |
43 Days Since On Sale |
50 Days Since On Sale |
57 Days Since On Sale |
64 Days Since On Sale |
AVERAGE
|
|
Issue 1 |
96
|
97
|
97
|
98
|
117
|
103
|
104
|
82
|
93
|
99 |
|
Issue 2 |
55
|
62
|
61
|
72
|
70
|
45
|
72
|
61
|
80
|
64 |
|
Issue 3 |
67
|
114
|
91
|
101
|
96
|
103
|
102
|
101
|
110
|
98 |
|
Issue 4 |
159
|
140
|
152
|
121
|
145
|
154
|
127
|
136
|
119
|
139 |
|
Issue 5 |
101
|
103
|
110
|
89
|
102
|
126
|
111
|
87
|
107
|
104 |
|
Issue 6 |
170
|
136
|
124
|
118
|
121
|
90
|
135
|
155
|
114
|
129 |
|
Issue 7 |
112
|
79
|
88
|
129
|
102
|
103
|
89
|
88
|
111
|
100 |
|
Issue 8 |
122
|
84
|
124
|
82
|
110
|
111
|
123
|
107
|
107
|
108 |
|
Issue 9 |
79
|
97
|
64
|
103
|
73
|
82
|
77
|
75
|
80
|
81 |
|
Issue 10
|
74
|
105
|
105
|
80
|
82
|
85
|
74
|
102
|
91
|
89 |
This example shows much the same as the
example of the weekly magazine. There are cases where an issue over performs in
respect to the average (e.g. Issue 4). Again, when an issue does better than
average it tends to do so over the majority of weeks in the field. In the case
of Issue 4, nine weeks out of nine produced higher than average indices.
Similarly, if an issue performs worse than average, Issue 9, it does so
consistently. In the case of Issue 9, eight measures out of nine consistently
underperformed from average.
c. Approaches
to Modeling
As is described in detail elsewhere,
the input data for the estimation process to determine the relative
issue-to-issue is based on successive week samples of approximately 5,000
persons. For weeklies there are 8
successive weekly samples for each issue and for monthlies there are 16
successive weekly samples. The basic
sample of 5,000 is weighted on the basis of all demographic variables each
week.
For the ith magazine, the
relative weekly measurement RWM for the kth relative week across all
of the issues (j=1,…J) extracted. A
mean, standard deviation and coefficient of variation are computed for the each
of the K relative weeks. Let CV(i,k)
denote the observed coefficient of variation for the kth relative
week of the ith magazine.
For the kth relative week of
the ith magazine, we compute the probable coefficient of variation
PCV(i,k). This probable coefficient of
variation is computed under the assumption that there is “no variation” in the
actual audience of the ith magazine at the kth relative
week of issue life.
Finally by subtracting the probable
coefficient of variation from observed coefficient of variation we obtain
ISCV(i,k) the issue specific coefficient of variation for the kth
relative week of the ith magazine.
These issue specific coefficients of variation are averaged over the k
included weeks. This average produces
ISCV(i) the issue specific coefficient of variation for the ith
magazine.
Within each of the k relative weeks
each of the means for the various issues are converted into indicies. These indicies are averaged across the
various included weeks to produce an initial issue specific index value. Let INDX(i,j) denote the initial average
index for the jth issue of the ith magazine. Using a power transformation SINDX(i,j) =
INDX(i,j)p, we solve for p so that the coefficient of variation of
the resulting SINDX(i,j) values is equal to ISCV(i)[9]
d. The
MRI National Study and the Issue Specific Study
As asserted above, the Issue Specific
Study is not intended to replace the MRI National Study. Rather, the two
studies work together. In effect, the Issue Specific Study introduces the
variability from one issue to another using the average issue audience from the
National Study as the base to demonstrate that variability.
Average issue data by title from the
National Study are used as references for the Issue Specific data. Once an
issue is tied to a reference National Study, that reference is never changed.
For example, all issues released in the first Issue Specific release (June 19,
2007) will be referenced to the Spring 2007 National Study. Once the Fall 2007
National Study was released, it replaced the Spring 2007 as the reference.
Every issue released in the Issue Specific Study subsequent to the Fall 2007
release was referenced to the Fall 2007 National Study. Those issues that were
released prior to the release of the Fall 2007 will always remain referenced to
the Spring 2007.
From the Issue Specific Study, MRI
establishes an index[10] for each issue of
every released title. This index represents the performance of that given issue
in respect to other issues of that title (indices substantially over 100 are
those that perform better than average, those substantially below 100 are those
that perform worse than average, and those around 100 perform about average).
In general, a twelve-month aggregation
of issue specific data will serve as the reference baseline for the initial and
subsequent six months of measured issues in the Issue Specific Study. Every six
months, consistent with the release of the new MRI Fall or Spring National
report, a new baseline is created by dropping the oldest six months of issue
specific data and replacing them with the most recent six months of collected
issue specific data. Therefore, MRI generates issue indices against a baseline
of data collected during the same period of interviewing in the National Study.
Multiplying the index and the average
issue audience from the National Study produces the issue audience for each
issue (this is done for total adult, male and female using the average issue
audiences from the National Study for each of the three groups).
For example, a hypothetical weekly
magazine (MAGAZINE A) has a 20 million audience in the Spring 2007 National
Study. Let’s assume this magazine had an issue (ISSUE 1) that over performed
and produced an index of 120 in the Issue Specific Study. The issue audience
would be calculated in the following way:[11]
|
|
Average Issue Aud MRI SPRING 2007 |
Issue Specific
Index |
Issue Audience |
|
ISSUE 1 |
20,000,000 |
120 |
24,000,000 |
X =
e. Deliverable
The deliverable is provided online.
Users are given username and password information in order to log into the
online spreadsheet. The spreadsheet includes indices for adult, male and
female, age, income and education. Additionally, the spreadsheet will publish
the issue audience for each released issue (the calculation of which is
demonstrated above in the box).
All indices come directly from the
Issue Specific Study. Male and female issue audiences are calculated using
these indices. The total adult issue audience is calculated by summing up the
male and female issue audience. Finally, the total adult index for an issue is
calculated by dividing the total adult issue audience by the total adult
average issue audience figure from the MRI National Study.
In order to enhance the data provided,
each magazine cover is linked within the spreadsheet as a reference for the
user.
The first release included data for
total adults, male and female for each released issue by title. Subsequent to
the first release of this study, MRI released additional demographic
information. Age information for each released issue by title was released as
of July 2008, education information was released in August 2008, and income
information was released in September 2008.
III.
Disclaimers
(1)
Non-responding and non-reporting persons may have media habits which
differ from those of respondents. Therefore, non-responding persons and
other limitations in the original sample may create additional bias in data.
(2) The sample, sample design and /or
response patterns may preclude proportional representation of certain groups
within the population such as ethnic groups, racial groups, persons in certain
income or education groups or persons whose primary language is other than English.
Such persons may have media habits which differ from other persons.
(3) Human and computer processing
errors may occur before or after Mediamark Research & Intelligence receives
the survey data.
(4) Defects and limitations found in
data and covers supplied by others (e.g., Survey Sampling Inc., e-Rewards,
clients sending covers, etc.) are inherent in MRI estimates based thereon.
Appendix 1:
Modeling/Release Standards
DATE: June 11,
2009
There are two levels of release
evaluated for the Issue Specific Study. First, does a magazine meet the requirements for release? Second, if the magazine
is released, do all the issues of
the magazine meet the requirements for release? There is an interplay between
these two levels of assessment in that, for example, if not enough data are
present to release a minimum number of issues of a potentially released title,
that title will not be released. The release standards are as follows:
1. As
of June 11, 2009,
MRI will use Spring 2009 as the reference period for all
magazines. If a magazine is not released in the Spring 2009 report, it cannot
be released in the Issue Specific study.
2. MRI
cannot include magazines in the service if they have less than 4 issues that
can be delivered at the reporting date.
3. MRI
must remove all known split cover issues that were not handled properly,
frequency change magazines and magazines with inconsistency execution of split
covers from reporting.[12]
4. Magazines
not released as of June 11, 2009 (Spring 2009 release) are not available for
release review until December 2009 (with the Fall 2009 release).
If a magazine meets initial
requirements for release, its data will be modeled for release. Certain
principles guide the modeling of the Issue Specific data:
1. Each
issue must have a minimum 6 weeks of measurements for weeklies
2. Each issue must have a minimum 8 weeks of measurements for bi-weeklies, tri-weeklies, monthlies and bi-monthlies
3. For
weeklies, 42 days are used as a maximum period for comparison.
4. For
bi-weeklies, 56 days are used as a maximum period for comparison.
5. For
tri-weeklies, 63 days are used as a maximum period for comparison.
6. For
monthlies, 63 days are used as a maximum period for comparison.
7. For
bi-monthlies, 84 days are used as a maximum period for comparison.
8. MRI
cannot model later weeks for an issue; only recent issues that have met a days
old requirement completely (no modeled numbers at the end) can be included.
9. MRI will trim all indices at 72 and
140.
If an issue of a released title
violates any of these modeling rules, it cannot be included in the model (and
is thereby not released). This, in turn, may have implications for the release
of the title (if too many of the issues have to be suppressed for violating
these rules, for example).
Appendix 2:
Copy of Invitations to Panelists
SSI Invitation 1: This is the invitation emailed to
potential respondents 25 years of age or older.
Thank you for agreeing to participate
in this media behavior study.
Click below to take part in this
research study and play the new SurveySpot Instant Win game. You’ll also be entered
into our $25,000 sweepstakes. If you have questions about this survey, please
reference survey number 668068.
Please be assured that all
responses to this survey will remain completely confidential and will be used
only in combination with all other responses received.
This study should take approximately 10
minutes of your time.
Please be completely honest, there are
no right or wrong answers
To record responses, select the box or boxes
that correspond to the answer choice. To advance the survey to the next page,
use the button at the bottom of the page marked NEXT.
During the course of the survey, please
do not use the “BACK” button that is built into the web browser.
Click here to begin:
Please answer each question in order.
Thank you.
SSI Invitation 2: This is the invitation emailed to
potential respondents between the ages of 18 and 24.
Thank
you for agreeing to participate in this media behavior study.
If you qualify and complete this
survey, you'll receive $3. In addition, you can play the new SurveySpot Instant
Win game and be entered into our $25,000 sweepstakes. If you have questions about this survey, please
reference survey number 668068.
Please be assured that all
responses to this survey will remain completely confidential and will be used
only in combination with all other responses received.
This study should take approximately 10
minutes of your time.
Please be completely honest, there are
no right or wrong answers
To record responses, select the box or
boxes that correspond to the answer choice. To advance the survey to the next
page, use the button at the bottom of the page marked NEXT.
During the course of the survey, please
do not use the “BACK” button that is built into the web browser.
Click here to begin:
Please answer each question in order.
Important: we expect this survey
to take 12 minutes. Please take your time and read each question carefully.
Sometimes, if interviews are rushed through they can’t be used. As a result you may not be eligible for the
survey incentive or reward.
Thank you.
e-Rewards Invitation:
Subject:
Get Rewarded For Your Time - Study about Media Behavior
Dear
Wave One,
Based
on your e-Rewards profile, you are invited to earn e-Rewards Currency for
participating in a research survey. If you qualify and complete the
survey:
Full
reward amount: $2.50 in e-Rewards Currency
Full
survey length: approximately 10-12 minutes
To
complete the survey and earn e-Rewards Currency, simply click the link below,
or copy the URL into your browser:
We
encourage you to respond quickly -- this e-Rewards invitation will be available
only until a predetermined number of responses have been received. Please
Note: you will only receive e-Rewards credit for taking the survey once.
Continue
to check your inbox and your Member home page for future opportunities to earn
e-Rewards Currency.
We
value your time,
The
e-Rewards Team
Skiing is published seven times a year.
Issues for this title are not published between May and August. One of the
Issue Specific release rules indicates that a title will be withheld from
release if the reference period for the issue specific data does not include
data that represents the entire year.. This rule was established to ensure that
titles have representation in all seasons in its reference. Because Skiing does
not publish in the summer months, it is a special case. Therefore, Issue
Specific data for Skiing will be released.