# Help with research paper! | Research Methods 2 |

I have already added the title page and the data needed which will be attached. I will also be attaching the procedure section so you are aware of what the study is about. I will NOT be using a t test, I ran two anova’s and a chi test!

3. Methods Section: I expect the following format **(15 points)**:

a. For this paper, the methods section starts on page 2.

b. Write **Method **at the top of this page, make it bold, and center it (see the top of this page as an example!)

c. The participants section comes next. The word **Participants **is bolded and left justified. In this section …

i. Tell me who your participants were (college students, family members, friends?) and how many there were.

1. Note: If a number starts a sentence, then spell out the number. That is, “Two-hundred and five participants participated in this study.”

2. If a number is mid-sentence, you can use numerals. “There were 205 participants in this study.”

3. But keep numbers consistent. If you spell out a number at the start of the sentence, carry that through and spell out other numbers in the sentence.

4. For statistics, always use numbers (for the mean, SD, %, etc.)

ii. Provide frequencies and descriptive statistics for relevant demographics.

1. For some variables—like ethnicity and gender—you only need to provide frequency information (the number of participants who fit that category). “There were 100 men (49%) and 105 women (51%) in the study.” Or “The sample was 49% male (*N*= 100) and 51% female (*N*= 105).”

2. Other variables—like age—are continuous (rather than categorical), so use descriptive statistics here (the range, mean, and the standard deviation). “Participants ranged in age from 18 to 77 (*M*= 24, *SD*= 3.50).” or “The average age of participants was 24 (*SD*= 3.50).” Your TA can help you find the mean and standard deviation for this assignment, though information is also available in a lab powerpoint.

3. Make sure to italicize the *N*, *M*, and *SD*(the letters, not the numbers)

d. **Materials and Procedure**

i. For this section, things are flexible. Some studies include Materials and Procedure in the same section while others break them up into two sections. This is a matter of choice.

1. In general, the more complex the design, the better it is to split up the methods and results. In one section, the author may describe the materials; in the next, they describe what participants did with those materials (the procedure). This is one option for you. However …

2. However, your “Paper II: Methods, Results and Discussion (Study One)” is simple enough that I strongly recommend combining them into __ one__overall Materials and Procedure section.

ii. Again, the words **Materials andProcedure**are flush left. In this section …

1. Provide information about your materials and your procedure.

a. I suggest starting with your procedure. Tell your reader what your participants did in the order participants did them. Be specific here. I have the following recommendations:

i. First, talk about the oral informed consent procedure.

ii. Second, talk about the three versions of the SexualityPriming questionnaire. __Provide enough detail so that your readers know how the three conditions differ. As a reader, I need to able to replicate your design, so you need to give me enough detail so I can do so.__(Hint: Copy and paste the various questions or refer the reader to an appendix that has those materials!)

iii. Third, talk about your dependent variables (that is, your survey questions. For these dependent variables, once again provide enough detail so I know __exactly__what questions you asked. For example, “Participants provided their gender, age, and race”. For other dependent variables, tell me how the responses were recorded (yes/no, true/false, a scale of 1 to 6, etc.). If you used a scale, note the endpoints. That is, does a 1 mean it is high or is it low? “Participants were asked, ‘How frustrating was this task?’, and they responded on a scale from 1 (very frustrating) to 9 (not at all frustrating).’” Your study has a few really important DVs (including several DVs about their impressions of Riley Washington, the Facebook user, on elements of sexuality and romance). For these DVs, you once again need to tell me what they are __specifically__!

iv. Fourth, make sure to highlight which specific DVs you analyzed. If there are DVs participants completed but you did not analyze it, feel free to say those DVs were not analyzed, but if you analyze them in the results section, then be specific about them in the methods section.

v. Finally, mention debriefing

e. There is no set minimum or maximum on the length of the methods section, but I would expect __at least a page or two__(though probably more. After all, your own research script took up several pages – you should provide a similar level of depth and detail in your methods section!). Missing important aspects of your IVs and DVs or presenting them in a confused manner will lower your score in this section.

f. Remember, make sure that another researcher can replicate your study based on your methods section. If they can’t, then you may not have enough detail!

4. Results Section: I expect the following format **(10 points)**:

a. The results are the hardest part of this paper, and your lab powerpoints will help you with this part of the paper (also refer to the crash course statistics quizzes, which walk you through similar analyses!).

b. First, write **Results**at the top of this section, center it, and use boldface. This section comes directly at the end of the methods section, so the results section DOES NOT start on its own page.

c. For this assignment, include statistics about the most important variables in your study, including your IV (Priming Condition –Sexuality, Romance, and Education) and the DVs you feel are most important to your hypotheses. I suggest looking at ONE independent variable related to the sexuality category (flirtatious, seductive, sex, or provocative)as well as ONE independent variable related to the romance category (sensitive, kind, tender, or sentimental). Note that some instructors may not do this Sexuality Priming study at all, but the results section should follow the same guidelines regardless of your study topic.

d. For this paper, **you must run at least three different analyses on three** **different****dependent variables**. One __must__be a chi square for the question asking participants which to recall the theme of the advertisements (our manipulation check, which looks at the three options for the nominal variable in Part VI). At least one of the remaining two analyses must be a One Way ANOVA (I actually recommend that both of your last two analyses focus on One Way ANOVAs). The third analysis can be either an ANOVA or a *t*-Test. Since all ten of the Riley Washington impression questions in Part II are scaled 1 to 6, I recommend running ANOVAs on two of those ten dependent variables (on in the sexual category and one in the romance condition). Now, you could run an ANOVA on the question “Riley seems sexy” OR you could run a *t*-Test on the question “Riley seems sexy”, but because it is the same dependent variable, that only counts as __one__DV. We count the number of DVs you analyze – NOT the number of statistical tests you run!

i. __ Chi square__: Your first analysis will be a chi square, which you use if your DV is categorical (yes / no; yes / no / maybe; male / female, or … in our case, we have our “Theme” based questions in Part IV (The theme involved sexuality, romance, or education). So let’s discuss the chi square, which does not look at means but rather counts how many responses there are compared to how many you would expect.

1. Consider the DV in Part VI of your questionnaire – “Without looking back, tell me the general theme of the advertisements you saw on the prior page: They focused on sexuality, They focused on romance, or They focused on education” Here, you can run a chi square looking at the frequencies of the three answer options

2. We are interested in the chi square (*χ*2) and *p*value. We also provide percentages for each of our groups (rather than means and *SD*).

a. “Using the priming condition as our independent variable (Sexuality, Romance, or Education) and the general ad theme participants recalled seeing as the dependent variable, we saw a significant effect, *χ*2(4) = 68.49, *p*< .001. Most participants in the sexuality condition recalled see an ad theme based on sexuality (98%); most participants in the romance condition recalled see ads about romance (96%); and most participants in education condition recalled seeing ads about education (90%). This indicates that participants saw our manipulation as intended.”

b. Alternatively, you can just look at correct versus incorrect responses. This is a bit trickier to run in SPSS, since you need to add up all those who correctly remembered the ad (those in the sexuality condition who recalled the sexuality theme + those in the romance condition who recalled the romance theme + those in the education who recalled the education theme) and compare them to people who recalled an incorrect theme. In this instance, you wouldn’t want the chi square to be significant. That is, you might conclude that *χ*2(4) = 1.49,*p*> .05, indicating that there was no difference between those who got the theme correct across the three different conditions. (In other words, participants weren’t more correct in one condition compared to another). My advice is to go with the chi square in a. above

c. Make sure to italicize the *χ*and *p*

ii. __ ANOVA__: Since you have a condition independent variable with three levels (e.g. Sexuality, Romance, or Education), the most appropriate test is a One-Way ANOVA if your DV is scaled (like a 0 to 6 scale or a 1 to 6 scale). Your lab and lecture powerpoints show you how to conduct an ANOVA, but there are some guidelines I want to give you about how to write your results. Below, I am going to walk you through one analysis specific to this paper. However, keep in mind that you can run ANOVAs on several different DVs.

1. First, there are several dependent variables to choose from. For my example analysis below, I want to focus on Part II in your survey (impressions of Riley Washington). Since each of the ten questions in Part II are scaled variables that range from 1 to 6, each uses an interval scale, which is perfect for an ANOVA. (Other questions we can look at are all of those ranging from 1 to 6 in Part III).

2. Second, given that this study has one IV with three levels and one DV that is on a continuous (ratio or interval) scale, a __One-Way__ANOVA is the best test to use to see if there are significant differences among the levels. We look first at the ANOVA table (or *F*table) and focus on the between subject factor. We note the degrees of freedom, the *F*value itself, and the *p*value. (We’ll get into two-way ANOVAs later in this course, but here we only have one independent variable, so it is a one-way ANOVA. Yes, we have three levels to our IV, but it is still only one IV).

3. If the *p*value is significant (less than .05), we have one more step to take. Since this is a three level IV, we need to compare mean A to mean B, mean A to mean C, and mean B to mean C. We do this using a post hoc test (try using Tukey!). That will tell us which of the means differ significantly. You then write up the results. For example, let’s say I ran an ANOVA on the dependent variable “Riley seems sexy”. My write up would look like this (though note: I completely made up the data below, so don’t copy the numbers!) …

a. “Using the priming condition (Sexuality v. Romance v. Education) as our independent variable and ratings of “Riley seems sexy” as the dependent variable, we found a significant condition effect, *F*(2, 203) = 4.32, *p*< .05. Tukey post hoc tests showed that participants thought Riley seemed more sexy in the sexual condition (*M*= 4.56, *SD*= 1.21) than participants in both the romance (*M*= 2.24, *SD*= 0.89) and education (*M*= 2.23, *SD*= 0.77) conditions. The romance and education conditions, however, did not differ from each other. This supports our prediction that participants exposed to aggressive advertisements are more likely to rate Riley high in sexuality than those who are exposed to ads about romance or education.”

i. Note there are lots of possible outcomes. The one above essentially says that condition S (Sexuality) differed from R (Romance) and E (Education), but that R and E did not differ from each other (In other words, S ≠ R = E). However, we might also find that NONE of the three conditions differ from each other (S = R = E) or we might find that ALL conditions differ from each other (S ≠ R ≠ E).

ii. As an example for this latter (S ≠ R ≠ E), I would predict no differences between the three conditions for the dependent variables “Riley seems educated” and “Riley seems outgoing”

b. Make sure to italicize the *F*, *p*, *M*, and *SD*(as in the example)

c. Pretty simple, right! I suggest going back and doing this same procedure for at least one additional scaled DV (like questions 3, 8, or 10 in Part III).

d. However, if you choose you can do a *t*-Test on one of those other dependent variables as well. Here’s how:

iii. **t**__ -Test__: If you have only two levels to your IV (e.g. Sexuality or Education only), things are even more simple.

1. Here, you will run a *t*-Test (a *t*-Test looks at differences between only two groups). Again, your lab presentations tell you how to run this, but you can do it on your own as well (you can even run this if your study originally has three levels to the IV – when you go into the *t*-Test menu in SPSS, choose “define groups” and select 1 and 3 (Sexuality = 1and Education = 3). This will let you look at two of the groups! You could also select “2 and 3” or “1 and 2” where the Romance condition = 2).

2. Rather than an *F*value, we will look at the *t*value in the *t*-Test data output. Here, we have one number for the degree of freedom, we have the *t*value, and we have the *p*value.

3. The nice thing about a *t*-Test is that since you only have two groups, you do not need a post hoc test like Tukey (you only need that if you have to compare three means. Here, we only have two means, so we can just look at them and see which one is higher and which is lower when our *t*-Test is significant). Then just write it up …

a. “Using the priming condition (Sexuality v. Education) as our independent variable and ratings of “Riley seems sexy” as our dependent variable, we found a significant condition effect, *t*(203) = 8.12, *p*< .05. Participants rated Riley as more sexy in the sexuality condition (*M*= 5.56, *SD*= 1.21) than participants in the education condition (*M*= 2.23,*SD*= 0.77).”

b. Repeat for other dependent variables

c. Make sure to italicize the *t*, *p*, *M*, and *SD*(as in the example)

iv. Statistics order recommendation: For this paper, start your results section with the chi square (your manipulation check). Then talk about your main analyses (Riley impression questions). Make sure the analyses line up with your hypotheses.

e. There is no page minimum or maximum for the results section, though I would expect it to be at least a paragraph or two for __each__dependent variable

5. Appendices**(4 points)**

a. I want to make sure you are including the correct numbers in your results section, so I want you to include all relevant SPSS tables for each of your analyses in a series of appendices.

i. Appendix A: Include your tables for age, gender, and ethnicity.

ii. Appendix B: Include your tables for your chi square and the crosstabs

iii. Appendix C: Include your tables for your first dependent variable (This must be an ANOVA table, the descriptive statistics table for that ANOVA, and the post hoc test whether it is significant or not)

iv. Appendix D: Include your tables for you second dependent variable (Although I prefer a second ANOVA like iii. above, you could include *t*-Test tables here. This would involve both the descriptives for the *t*-Test and the *t*-Test output itself

b. Hint: The best way to get these tables is to copy them directly from SPSS. In the SPSS output, right click on the table, copy it, and then paste it into your appendix. Another alternative is to use a “snipping” tool (search “snipping tool” in Microsoft Word to find it). You can highlight an area on any computer page and save it as a picture. Copy the picture and paste it into your appendix. Easy!

6. Discussion Study One **(2 points)**

a. In this section, tell me about your findings and if they did or did not support your results. It might help to refer back to your hypotheses “We expected to find A but instead found B” or “We expected to find A and our results supported this hypothesis.” Explain using plain English why you think your study turned out the way it did.

b. IMPORTANT – Do NOT give me statistics again here. I can find those in your results section. Here, all I want is a plain English summary of your findings.

c. Also, don’t give me results for a DV if you did not run an analysis on that DV. Only tell me about the results you actually looked at in the results section.

d. There is no length requirement for this section, but I recommend at least four or five sentences

7. Overall writing quality **(3 points)**

**a.** Make sure you check your paper for proper spelling and grammar. The FIU writing center is available if you want someone to look over your paper (an extra eye is always good!) and give you advice. I highly recommend them, as writing quality will become even more important on future papers. I also recommend visiting the FIU Research Methods Help Center if you need additional guidance with writing or statistical analyses. Also, remember to upload this paper through the Pearson writer before uploading to blackboard!