PSYC5252 - Assessment 3

Week Due Sunday 14 August at 11.59pm AEST

Being a research consumer You will adopt the role of a journal reviewer. You will be required to critically review the methodological and data analysis components of a mock journal article, as well as re-analyse the data and present it in correct APA style.

This assessment is very different from any assessment that you would have done in the past. It has been deliberately designed for this course, and to complement and test your knowledge of research methods and data analysis.

In this lab, you are cast in the role of someone who is reviewing a journal article that has been submitted for publication. When manuscripts are submitted for publication, they are reviewed by at least two people using a system called blind peer reviewing. What this means is that you don't know who is reviewing your paper, and your name and affiliation is taken off the manuscript so the reviewer does not know the author's name or where the paper comes from. This is an attempt to avoid all sorts of biases creeping into the review process.

The mock manuscript is a single file that can be found on the Moodle class space. I have tried to present it as closely as possible as a "real" manuscript. It is obviously more basic than the research you would find in a manuscript presented for publication, and the citations are made up (no references have been provided; do you think that is something you should comment on?), but the basic structure, format, and presentation is pretty much what you would see in a "real" manuscript.

What is unusual is that you will be provided with the raw data in SPSS format that formed the basis for the manuscript. This is rare, but it does happen. Authors are regularly asked to keep copies of their raw data so that the article reviewer or journal editor can check the statistical analysis. Making raw data publicly available for quality control is starting to become more common. This is part of what is referred to as the Open Science movement. There is a good short introductory article on open science from The Conversation here: https://theconversation.com/research-transparency-5-questions-about-open-science-answered-76851. Finally, there is a letter to the editor that will form the basis for your report.

Your job is to fill in the section under the heading "Specific points of note". What you need to do is point out the flaws in background, design, method, analysis, and interpretation that has led to your strong rejection of the paper. You need to highlight what the author has done and where he has gone wrong, and you need to correct his errors by providing the correct analysis of the data, including correct versions of the tables, if need be. Your focus must be on the methodological aspects of the manuscript; for example, you shouldn't waste space making comments on the APA style of the paper, or saying that the discussion is too short, or stuff such as that. I repeat--your focus should be on the design, analysis, interpretation, and method. That is the main reason why you have rejected the paper. You certainly aren't meant to criticise the actual research; in other words, there's no need to provide reference material on the actual topic. It's the methodological issues that count. By the way, this also includes ethics.

There is no set format for your review, just as there is no single perfect format for the way "real" reviewers approach reviewing manuscripts. You can use numbered or bullet points, or you can use sub-headings. In fact, part of the assessment will be based on the structure you apply to the task and how well that gets your message across.

This is a totally new assessment for you, and you may feel that you are swimming in the dark when it comes to presenting the report. This is where the Discussion Board comes into its own. Please use it freely to post any questions you might have, no matter how naive you think they might be.

The word limit for the report is 2000 words; that's 2000 words of what YOU write, so it doesn't include the stuff that I have provided for you in the mock letter to the editor. The 2000 words does not include material in tables or figures, or references.

Assignment FAQ

Q. How should I format the report?
A. There isn’t really one correct way of formatting the report. The most common approach is to itemise the errors in the first section, and then provide corrections to those in a following section. However, that’s just one approach. You could also highlight an error and then provide the correction before moving on to the next error and correction. That’s also fine. This isn’t an APA style research report. You are not writing a report. You are reviewing a manuscript. It’s a different exercise. The format itself is not as important as making sure you have correctly identified errors and corrected them.

Q. Should I use references? If so, how many?
A. This is not really a report that requires extensive referencing. In fact, it is possible to write this report with no references at all. This is all about assessing your knowledge and understanding of the research methods that we have been discussing in this subject. However, if you make a claim that you think requires some support, then by all means include a supporting reference. For example, you might say something along the lines of “It is generally argued that minor deviations from normality are not important, particularly when the sample size is greater than 50 (Smith, 2017)”. That’s the sort of statement that would benefit from a supporting reference. When it comes to references, I always follow one golden rule—when in doubt, pop in a reference. There are certain phrases that almost seem to demand a reference—“It has been argued”, “It is generally accepted”, “Researchers have claimed”, and so on. These are the sorts of phrases that really seem to suggest a reference. In terms of the number of references, I never set a number. To me, it’s all about quality over quantity. Two good references can be better than 20 useless ones.

Q. I’m seeing errors in the paper apart from errors in methodology and the other things you listed. Should I highlight those?
A. There’s no point in doing so; they will be ignored in assessing your report. Stick to the areas that I’ve highlighted in the body of this guide.

Q. Is the data real?
A. No, the data has been made up for this assignment.

Q. My assignment doesn’t come anywhere near 2000 words. Is that a problem?
A. No, not at all. There is no minimum word count. 2000 words is the absolute maximum. There NO 10% wriggle-room on that. 2000 means 2000, and that’s it.

Q. What needs to be in APA style?
A. Any tables and graphs that you present must be in properly formatted APA style. In addition, if you present any statistical results in text, that material must be formatted in correct APA style.

Q. Do I need to attach my own analysis of the results as an appendix?
A. No, you do not. You only need to submit one document—your review of the manuscript for Professor Mouse.

The Causal Relationships Among Sex, Television Viewing and Overweight in Young
Adolescents

The issue of adolescent obesity has been identified as one of the most important
public health issues facing Western societies. The most reliable survey data provides
convincing evidence that the proportion of adolescents who can be classified as obese (i.e.,
exhibiting a body mass index [BMI] > 30) has increased significantly over the last three
decades. In 1970, Pitt and Jolie found that only 6% of young American adolescents in the 12 –
14 age range met the criteria for being obese. By contrast, Schwarzenegger (2005) reported
that 15% of his sample of 1,500 young American adolescents exhibited a BMI > 30. Similar
trends have been reported in Australia (Rudd & Howard, 1999) and the UK (Blair &
Thatcher, 2003).
A range of causal agents have been proposed in an attempt to explain this trend,
including: the increased availability and affordability of energy-rich foods (Clinton, 1998), the
reduced presence of physical education in the primary and secondary school curricula (Bush,
2004), and the increase in passive forms of entertainment, such as computer games (Obama,
1998).
Television viewing, which is also a popular form of entertainment for young
adolescents, has also been targeted as a causal agent in adolescent obesity for a number of
reasons: first, it is passive; second, advertisers have been accused of targeting young
adolescents by saturating key adolescent viewing hours with advertisements for energy-rich
foods, particularly fast foods. However, television is not a new phenomenon, and there is
evidence that the amount of time that adolescents spend watching television has actually
decreased over the last 20 years (McGuire, 2003), although whether this can be explained by
the broader range of more recently available passive entertainment options for young
adolescents (e.g., online games, instant messaging, etc.) is open to speculation.,
Somewhat surprisingly, the relationship between television hours and obesity in
young adolescents has not been the subject of large-scale empirical investigations. There is
limited anecdotal evidence and considerable comment in the popular media regarding the
relationship between television viewing hours and obesity (see, for example, Jackson, 2002),
but very little reliable quantitative data exploring the association between these variables.
Certainly, there is no data exploring these relationships in Australian adolescents, and given
that there is cross-cultural evidence demonstrating notable differences in obesity levels and
other activities between Australian and American adolescents (Crowe, 2006), such an
investigation seems warranted. Further, there is some evidence for sex differences in both
obesity levels (Gaga, 1998) and television viewing habits (Cruise, 2004); hence, a valid
secondary question concerns whether sex moderates the relationship between obesity and
television viewing hours.
Therefore, the aim of the present study was to explore the causal relationship
between television viewing and overweight in a sample of young Australian adolescents.
Based on previous related research, the following hypothesis was tested: Will high levels of
television viewing directly cause an increase in weight? Further, based on the consistent
findings of the research reviewed above, a secondary hypothesis was tested: Will there be a
causal relationship between sex and the relationship between television viewing hours and
overweight? Will a stronger relationship be evident for girls over boys?
Method
Participants
A total of 50 young adolescents (30 male and 30 female) participated in the study.
The participants were recruited from three government secondary schools in the northern
region of Melbourne. Because the relevant classroom teachers provided permission for the
researcher to access the students in their class, and because the participants were under the
age of 18, neither parental consent nor consent from the adolescent participants was obtained.
The participants were told that their participation was compulsory; any student who insisted
on not participating was required to write an essay on the value of participating in
psychological research. The participants were randomly selected from the total school
population, with all students with family names beginning with the letter A – K instructed to
participate.
The mean age of the total sample was 13.4 years ( SD = 2.5 years) with boys ( M =
13.9 years, SD = 2.6 years) statistically significantly older than girls ( M = 13.3 years, SD = 2.2
years), as assessed using  2 .
Materials and Procedure
All participants completed a survey, purpose-built for the present study, that asked
them to report on their television viewing habits over the previous week. Participants were
asked to recall precisely what television shows they had watched over the last seven days. All
participants were given a copy of a relevant television guide to aid their recall. This survey
was completed in regular class time. Participants were asked to place their names on the
surveys so their data could be matched with the height and weight measurements that
followed. The completed surveys were then randomly distributed among the class members for
scoring. The survey materials had excellent psychometric properties with well-established
reliability and validity.
All participants were weighed in their classrooms while completing their surveys
using a set of commercial bathroom scales, and their height was assessed, barefoot, using a
simple tape measure. Both of these measures were taken by the other students in the
classroom as part of a work experience exercise. Based on these measures, each participant’s
BMI was calculated. This was done by the students in the class as part of a classroom
mathematics exercise.
At the end of the study, all students were given $20 McDonald’s vouchers as
payment for their participation.
Results
Data Coding and Exploratory Data Analysis
All data were entered into a single SPSS data file for analysis. Seven variables were
coded:
1. Sex
2. Total amount of television viewing hours over the previous week
3. Height in metres
4. Weight in kilograms
5. BMI
All continuous variables were initially analysed using exploratory data analysis
techniques to (a) check for data entry errors, and (b) test the assumptions of normality and
homogeneity of variance. Techniques included identifying data entry errors via visual
inspection of histograms, stem-and-leaf plots and normality plots; inferential tests for
normality, such as Levene’s test; and both the Kolmogorov-Smirnov and Shapiro-Wilks
tests for homogeneity of variance.
All continuous variables exhibited high levels of positive and negative skew, with p =
.000 for all normality tests. A range of transformations were attempted, however, none
resulted in any notable improvement in the distributions of scores, so the original data
were used for all subsequent analyses. Further, the homogeneity of variance assumption,
using sex as the nominal variable, was found to be violated at p > .50 for all outcome
measures. A reciprocal log-root transformation was applied to the sex variable, and
subsequent testing found that the assumption had been met.
Table 1 presents the descriptive results for all variables. These results clearly
demonstrate strong and statistically significant sex differences in all measured variables:
height, weight, BMI and television viewing hours. Females had significantly higher BMI’s
than males, and watched significantly more television, clearly supporting the hypothesis
that being female is causally related to both a increased BMI and watching more television.
The overall correlations, and the correlations for males and females separately are
shown in Table 2. For the overall sample, there was a strong and significant positive
correlation between both BMI and weight, and television viewing hours. For all
correlations, the relationship between television viewing hours and the two weight
measures, weight and BMI, was statistically significantly stronger for males than it was for
females. These correlations were all assessed using  2 analysis, which allows causal
conclusions to be reached.
Regarding sex differences, a series of paired samples t-tests demonstrated significant
sex differences for all variables: height, t(3.33) = 34.71, p = .002; weight, t(38) = 1.79, p =
.081; television viewing hours, t(38) = 3.49, p = .069; and BMI, t(37.18) = 6.58, p = .52.
Examination of the descriptive statistics in Table 1 indicates that, on average, males were
taller than females, whereas females were significantly heavier, had higher BMI’s and
watched more television. All of these t-tests were based on unequal variance due to
violation of the homogeneity of variance assumption.

Discussion
The hypothesis that there would be a strong and significant causal relationship
between BMI, weight and television viewing hours was clearly supported. For the overall
sample, as well as for the male and female participants separately, there were strong and
significant correlations between the two weight-related measures and television viewing
hours. This result accords with previous related research, and supports the long-held
perception that television viewing, a passive form of entertainment for young adolescents,
is related to a propensity to overweight. Most notably, however, given the nature of the
research design, the present study is the first published investigation to demonstrate a
causal relationship between television viewing and overweight. This result has important
scientific and clinical applications in the area of adolescent obesity and strongly suggests
that an intervention program that focuses on limiting, or even eliminating, television
viewing, has a high probability of leading to substantial and permanent weight loss.
Further, there were clear sex differences in the nature of the relationship between
television viewing and BMI, with much stronger relationships evident for females. This
result clearly argues for targeted interventions toward adolescent females, as it is clear
from the results of the present study that the excessive television viewing is a more potent
causal agent for obesity in females than males. This result also implies that females are
engaging in less physical activity than males, an interpretation that is supported by the
significantly higher level of television viewing hours for the female participants.
Future research might choose to focus on evaluating well-designed interventions for
reducing television viewing hours

Table 1
Descriptive Statistics for Males (n = 25), Females (n = 27) and Total Sample
Male Female Total
Variable M SD M SD M SD
Height (cms) 172.83 10.93 167.70 8.11 177.95 11.15
Weight (kgs) 81.33 12.06 78.00 9.06 84.65 9.01
BMI 4.47 11.81 4.81 27.61 5.19 28.11
Viewing Hours 25.77 10.82 14.53 30.15 27.96 40.00

Table 2
Correlation Matrix of Variables for Males, Females and Overall
Overall (N = 40)
1 2 3 4
1 Height (cms) r - .51 .54 .25
p .11 .01 .007
2 Weight (kgs) r - .69 -.63
p .000 .000
3 BMI r - <.001
p .72
4 TV Viewing (hours) r -
p
Males (n = 20)
1 2 3 4
1 Height (cms) r - .49 -.72 -.73
p .16 .029 .08
2 Weight (kgs) r - .55 .81
p .011 .000
3 BMI r - .78
p .000
4 TV Viewing (hours) r -
p
Females (n = 20)
1 2 3 4
1 Height (cms) r - .87 -.39 -.39
p .01 .086 .09
2 Weight (kgs) r - .87 .69
p <.001 <.001
3 BMI r - .79
p .000
4 TV Viewing (hours) r -
p