1
00:00:00,390 --> 00:00:05,880
The next step is to construct the outline of your results section. And first of all, there are a few decisions

2
00:00:05,880 --> 00:00:07,280
that you need to make.

3
00:00:07,290 --> 00:00:13,020
So one thing you need to think about is whether you want to visualize key results in figures such as

4
00:00:13,020 --> 00:00:14,920
for example a bar graph.

5
00:00:15,150 --> 00:00:20,160
Another decision you need to make is whether you want to use tables, which can make sense if you have

6
00:00:20,160 --> 00:00:22,400
extensive amounts of information.

7
00:00:22,500 --> 00:00:27,720
And another thing to think about is how you want to structure your results section because results sections

8
00:00:27,720 --> 00:00:29,200
are highly individual.

9
00:00:29,250 --> 00:00:30,910
They really depend on your results.

10
00:00:30,930 --> 00:00:35,110
And so there aren't many guidelines on how you should structure it.

11
00:00:35,130 --> 00:00:40,440
So in this lecture I want to help you as best as I can by giving you some general guidelines for each

12
00:00:40,440 --> 00:00:41,890
of these points.

13
00:00:41,910 --> 00:00:47,280
So one thing that is worth thinking about is whether you want to depict your key results in a figure.

14
00:00:47,880 --> 00:00:53,910
So for example for the fictional study that I discussed a couple of times about how to treat depression

15
00:00:54,360 --> 00:01:01,050
we could include a bar graph that depicts the mean depression of the placebo group and the

16
00:01:01,050 --> 00:01:03,630
mean depression of the treatment group.

17
00:01:03,630 --> 00:01:07,950
However, in that case that wouldn't really be that necessary because if you mentioned in text that the

18
00:01:07,950 --> 00:01:09,680
mean was lower for the treatment group.

19
00:01:10,230 --> 00:01:14,850
And if you also mentioned the means that basically already gives all the information that the reader

20
00:01:14,850 --> 00:01:15,780
needs.

21
00:01:15,810 --> 00:01:20,160
However, suppose that we had four means or even more.

22
00:01:20,160 --> 00:01:24,560
Well, then it can get a lot more difficult to imagine the pattern of the results.

23
00:01:24,700 --> 00:01:29,250
And in that situation it can make sense to add a figure to make this more clear.

24
00:01:29,460 --> 00:01:34,200
So this is something you need to think about. If something is already clear without figure then you don't

25
00:01:34,200 --> 00:01:35,520
need to use a figure.

26
00:01:35,520 --> 00:01:41,520
But if a figure helps to clarify your results it's definitely a great tool to use. And it doesn't have

27
00:01:41,520 --> 00:01:42,510
to be a bar graph.

28
00:01:42,510 --> 00:01:48,650
It can also be a scatter plot or a heat map or whatever makes sense in your case.

29
00:01:48,660 --> 00:01:51,940
Another thing to think about is whether you want to add a table.

30
00:01:52,020 --> 00:01:58,020
So this for example is a table from one of my studies and I had the problem that I had to report a lot

31
00:01:58,020 --> 00:02:04,140
of parameter estimates and the lower bound and the upper bound of their confidence intervals, which

32
00:02:04,140 --> 00:02:05,510
is a lot of numbers.

33
00:02:05,640 --> 00:02:11,100
And so I decided to report these numbers in a table because if I put them all into text this would be

34
00:02:11,130 --> 00:02:11,770
unreadable.

35
00:02:11,790 --> 00:02:12,530
Right?

36
00:02:12,570 --> 00:02:17,670
So if you have a lot of information and especially numerical information it can make sense to add this

37
00:02:17,670 --> 00:02:23,340
information in a table and then you can just refer to this table in the text. And by the way the same

38
00:02:23,340 --> 00:02:24,260
is true for figures.

39
00:02:24,270 --> 00:02:27,830
If you use a figure, refer to it in the text.

40
00:02:27,870 --> 00:02:28,150
Okay.

41
00:02:28,170 --> 00:02:30,230
Next we need to think about the outline

42
00:02:30,240 --> 00:02:35,400
and so the general structure of your text. And as I said, there aren't really any clear rules for that

43
00:02:35,410 --> 00:02:37,950
because people have so different results.

44
00:02:37,950 --> 00:02:43,350
And if your results section is very short then it can make sense to write it in just one text with paragraphs

45
00:02:43,890 --> 00:02:49,590
but if it's quite extensive then it can make sense to have subheadings. Whatever you do, two parts you find

46
00:02:49,590 --> 00:02:55,230
very often in the results section is one part where you describe what you did before the analysis.

47
00:02:55,230 --> 00:03:02,160
So this could be the first paragraph or heading such as "pre processing". And then after that, people usually

48
00:03:02,160 --> 00:03:06,900
describe what their main analysis are and I want to give you some pointers for each of these two

49
00:03:06,900 --> 00:03:08,050
parts.

50
00:03:08,160 --> 00:03:10,220
So what could you have done before the analysis?

51
00:03:10,230 --> 00:03:15,600
Well, you could have, for example, excluded participants or observations. Maybe because they didn't complete

52
00:03:15,600 --> 00:03:18,940
your whole study and so you just couldn't use their data.

53
00:03:19,110 --> 00:03:20,940
Or maybe you transformed the data.

54
00:03:21,030 --> 00:03:23,040
Maybe you combined two variables.

55
00:03:23,040 --> 00:03:29,240
Maybe you recoded variables or maybe you applied some kind of smoothing. Another thing that may happen is

56
00:03:29,240 --> 00:03:34,040
that you have missing data and you fill them in with some kind of statistical technique.

57
00:03:34,190 --> 00:03:38,840
And if you have done so then it's definitely a good idea to describe that first and tell us how you

58
00:03:38,840 --> 00:03:39,670
did it.

59
00:03:39,680 --> 00:03:44,210
So start your results section with whatever you did before the main analyses.

60
00:03:44,210 --> 00:03:49,730
Then next, you describe your main analysis and I wanna give you a general recipe that I always use

61
00:03:49,730 --> 00:03:51,630
when I describe my main analyses.

62
00:03:51,920 --> 00:03:58,610
So I always start by describing the research question that belongs to this analysis and this is particularly

63
00:03:58,610 --> 00:04:03,290
important if you have several research questions because then it can get very confusing for the reader

64
00:04:03,580 --> 00:04:05,610
what you are actually trying to do there.

65
00:04:05,630 --> 00:04:11,020
So I always like to start with restating the research question that belongs to this analysis.

66
00:04:11,090 --> 00:04:14,660
Then after that I describe the analysis that I used.

67
00:04:14,690 --> 00:04:18,200
So for example, did you use a t test? Did you use an ANOVA?

68
00:04:18,200 --> 00:04:21,700
Did you use regression? Or did you use something completely different.

69
00:04:21,710 --> 00:04:28,170
What is it that you used? And then I describe the results and then after that I do nothing.

70
00:04:28,310 --> 00:04:33,950
I'm done. So especially if you restate your research question, it can get very tempting to also try to

71
00:04:33,950 --> 00:04:37,790
give an answer to that question but resist the temptation.

72
00:04:38,030 --> 00:04:43,410
It's okay to re-state your research question but you shouldn't answer it in this section yet. Okay.

73
00:04:43,430 --> 00:04:48,110
So these are some general pointers on how to structure your results section and how to create an outline

74
00:04:48,110 --> 00:04:49,340
for that section.

75
00:04:49,340 --> 00:04:51,440
Now let's turn to the writing in the next lecture.