Preparing the Results Section

 

        RM 0300 (UG25) 01 - Research Writing

        The Results section should contain the summary of the findings

        The Results section should contain the summary of the findings, including the results of statistical analyses. This section should be written in a form that is predictable.

        Report the statistical tests of your hypotheses in the order in which they were originally presented.

        Do not give in to the temptation to start with the finding you like best or end with one that supports your favorite hypothesis.

        Lead the reader through your analyses in the order that is most logical, not necessarily the one that is most exciting.

        Do not discuss which hypotheses were supported and which were not; those sentences belong in the Discussion section.

        This is not the place to concern yourself with variety of sentence construction.

         It is best to use the same sentence format for every result that involves the same type of statistic.

        Results and Discussion can be combined into one section.

        As a rule, students should have separate Results and Discussion sections unless directed otherwise by the instructor.

        State results in the order that corresponds to the order of the hypotheses as presented in the introduction.

        Statistics

        Whenever an effect is significant, report the direction of that effect.

        For t tests, F tests, and chi-square tests

        For t tests, F tests, and chi-square tests, include the numerical value obtained for the statistic, degrees of freedom, and probability level.

        Reporting of confidence levels is also strongly recommended.

        For F tests, also include the mean square of the error term (MSE).

        Indicate the direction of effect; for example, Group A scored significantly higher than Group B is better than the scores of the two groups were significantly different.

        Follow this information with the actual f-test result.

        You must report actual means and standard deviations (or some other measure of variability) whenever you report that means differed.

        These may be provided in a table.

        Students are usually required to report a bit more information than is included in journals.

        For example, after a f-test result, you must indicate whether it was one-tailed or two-tailed.

        Also, you will have to include statistical values for all of your results, even the nonsignificant ones.

        Please note that the word insignificant does not belong in your paper.

        That is not a technical term; it is an insult.

        If you mean that the result was not significant, you should use the word nonsignificant.

        No one will be insulted that way.

        Use nonsignificant rather than insignificant if an analysis does not yield an acceptable level of significance.

        You should report exact p values provided by statistical packages.

        The main thing to remember in report-ing probability is to use the less than symbol and the equals symbol appropriately.

        If the value comes from a table, use the less than symbol (e.g., p < .05); if it comes from a statistical package, use the equals symbol (e.g., p = .023).

        But note this exception: If the computer print-out indicates a probability level of .000, you should write p < .001.

        A p value cannot be zero, so .000 on a printout indicates that the prob-ability has been rounded off.

        The real probability might have been .0000071.

        If you drop the last zero in the value on your printout and replace it with a 1 and then claim that p < .001, you have made an accurate statement (.0000071 is less than .001).

        There are only a few ways to phrase statistical reports.

        Use only very recent journals for this exercise because the older journals might follow rules that are out of date.

        Analysis of variance results usually contain the abbreviation ANOVA.

        The rule for abbreviations is the same for the Results section as for the rest of the manuscript.

        Introduce the abbreviation in parentheses the first time you use the term, and then use only the abbreviation thereafter.

        If you do not use the term a second time in the manuscript, do not introduce the abbreviation at all.

        The abstract does not count as part of the manuscript for this rule.

        Introduce abbreviations in parentheses and use the abbreviations rather than the full term thereafter.

        Usually, the F test is reported for the ANOVA.

        Students have varying degrees of familiarity with ANOVA results.

        Undergraduates are likely to have experience with one-way analyses and analyses using two independent variables.

        Therefore, these will be the focus here.

        If you compared three or more means in a one-way analysis, report the results using the term one-way analysis of variance.

        Remember that a significant finding means that at least one mean was different.

        Because this lacks precision, authors usually do planned or post hoc tests on these means.

         Planned comparisons are reported as such.

        Post hoc tests are usually named (e.g., Tukey or Scheffe), and a significance level is targeted prior to the calculation.

        You are also expected to provide cell means, cell standard deviations, and an estimate of the pooled within-cell variance.

        These often appear on a table.

        Analyses with two independent variables require the reporting of main effects and interactions.

        The safest way to report these is either with both main effects preceding the interaction or with the interaction first.

        Do not report one main effect, the interaction, and then the other main effect.

        As with t tests, if you say that a main effect was significant (e.g., “The main effect of color was significant”), take the opportunity to say right at that time what the direction of the effect was (e.g.,

        “The main effect of color was significant, with the blue-pencil group scoring higher than the green, F. . .”).

        Interactions generally leave you two choices for phrasing:

        (a) “The interaction between age and instructional condition was significant, F ...” or (b) “The Age X Instructional Condition interaction was significant, FP... .’

        Capitalization rules are somewhat unexpected: main effects are lowercase, but interactions are capitalized.

        Do not be concerned that you have several sentences in a row that are the same in structure.

        The reader will not be put off by this but rather will appreciate the clarity.

        Correlation results require the correlation coefficient, 7, and the
p value.

        Reporting correlation results can be a preposition nightmare.

        These are acceptable statements of correlation:

        X correlated significantly with Y.

        X and Y were significantly correlated.

        The correlation between X and Y was significant.

        The correlation of X and Y was significant.

        Correlations among X, Y, and Z were computed.

        The correlation of X with Y was significant.

        When statistical analyses are used to compare means, provide all relevant means and standard deviations in the text or in a table, but not in both.

        Authors should to report effect size estimates as well as significance tests.

        A few authors routinely report a (eta squared) with F-test results: F(1, 115) = 623.16, MSE = Ol, p < .01, 1? = .83.

        There are a variety of effect size estimates to choose from, and your instructor will guide you.

        Tables and Graphs

        Tables should be used sparingly because they are inconvenient for readers and expensive to publish.

        Your professor may instruct you to present certain findings in tabular fashion to give you practice.

        The APA guide-lines are unique and complex for tables.

        Although some basic rules are presented here, if you prepare a table, consult the Publication Manual for further directions.

        If you do use a table, you must refer the reader to it by number and indicate what will be found there.

        Tables are good for presenting data when you want to be precise.

        The order for backmatter in your manuscript is References, Tables, Figure Captions, Figures.

        For a line graph, differentiate lines by differentiating plot points.

        Use clear open and solid circles and triangles as plot points.

        Lines should all be solid rather than dashed or dotted.

        For a bar graph, use simple shading techniques to distinguish between sets of bars.

         White (no shading) and black are preferable to grays and stripes.

        If you need a third shade, use diagonal stripes.

        Do not use color for graphs.

        Place a legend inside the graph area.

        Do not label the lines themselves in a line graph; explain the meanings of the shapes of the plot points or the shadings of the bars in the legend.

        As with tables, figures are numbered consecutively in the order in which they are referred to in the body of the manuscript.

        Write the figure number lightly with a pencil (e.g., Figure 1) on the top right front of the page containing each graph, and write the word top on the back of the page to show which way is up.

        Also write the manuscript page header on the back.

        There is no title except on the figure caption page.

        Refer to all tables and figures at least once in the body of the paper.

        Useful Rules

        1. Letter symbols (e.g., N, p) are italicized.

        2. Greek letters are not underlined or italicized.

        3. Letters that are abbreviations (e.g., M, SD) should be used only in parentheses. In the narrative, use the word (e.g., mean, standard deviation).

        4. Use the symbol for percent (%) whenever it is preceded by a numeral (e.g., 3%). ;

        5. Use spaces between symbols and within equations as if each term were a word (e.g., p = .05).

        6. Use numerals for 10 and above; use words for nine and below.

        Exceptions:

        a. Never begin a sentence with a numeral. Look up spellings for numbers in the dictionary and pay attention to hyphen use.

        b. Use numerals below 10 if they are grouped for comparison with numerals above 10 (e.g., 3 out of 14 trials).

        c. Use numerals below 10 in an abstract.

        7. Use metric units unless the nonmetric is more familiar (e.g., 3 * 5 cards). In this case, put the metric equivalent in parentheses.

        8. Use a zero before a decimal point when the value of a number is less than 1, unless it can never be more than 1 (e.g., levels of significance, proportions, correlation coefficients).

        9. Rounding off: Use two decimal places when reporting inferential Statistics and, in general, p values. For means, use two decimal places as long as relevant differences can be seen with two decimal places. Otherwise, try to rescale, for example, converting centime-ters to millimeters.

        10. Abbreviations for any measurement you are likely to need are listed in the Publication Manual. Note that most, but not all, abbreviations for units of measurement are neither capitalized nor followed by a period. Leave a space between the numeral and the abbreviated unit of measurement.

        11. The plural of analysis is analyses.

        12. Between is used for two things: Correlations were computed between two variables. Among is used for three or more: Correlations were computed among three variables.

        13. Word your sentences so that statistical results are not in parentheses. Many statistical results contain parentheses of their own (containing degrees of freedom, for example). Set off these statistical results with commas instead.

        14. Do not use mathematical symbols as if they were verbs in your sentences. For example, the following is incorrect: The number of boys = 17. You can use the word equals if you really must write a sentence that uses it. The word was is also nice here. Use the symbol for mathematical terms when those terms are inside parentheses.

        15.  Common fractions are expressed in words (e.g., one half of the sample, three fourths of the liquid), and others are expressed as numerals (e.g., 31.2 pastries).