Chapter 4: Writing the Methods Section

# Methods Goal 3: Analyzing the Data

The third goal of writing your Methods section, Analyzing the Data, is to overview the data analysis. To do this, the author explains how the data have been analyzed (without describing the results of that analysis). Accomplishing this goal provides a preview of the central pieces of the research, including, but not limited to, objectives, questions/hypotheses, procedures/methods, and main results. Authors also make arguments about the value of the reported work in an effort to justify the need for addressing the niche and any previous attempts to provide solutions to problems or gaps in the literature.

Sometimes writers outline their papers to help the reader understand the structure of what’s to come. In addition to describing the specific actions taken during the analysis of the data, this goal contains explanations of actions taken prior to the analysis, during which researchers prepare the data for analysis and interpretation. This goal might also include a description of measures taken to establish the credibility of the investigation (e.g., reporting reliability, noting limitations, or mentioning follow-up work).

Turn your attention to these excerpts from published writing. The bolded parts demonstrate the writer’s attempt to accomplish Goal 3 by explaining how the data was analyzed.

## Examples

• Initially, all main effects (year, planting date, cultivar, and harvest date), as well as all of their associated interactive combinations, were included within the statistical model. Based on previous experience (Gunsaulis et al., 2008; Coblentz and Walgenbach 2010), management main effects (planting date, cultivar, and harvest date) were likely to interact with each other, and also to interact strongly with year. This hypothesis proved to be correct; regardless of the response variable (canopy height, growth stage, DM concentration, and yields of DM), nearly all interactive combinations of main effects were highly significant (P less than 0.01). To properly assess the DM yield potential of fall grown oat cultivars, an understanding these interactions was imperative, particularly those with environment (year); therefore, year was considered to be a fixed, rather than random effect. Furthermore, some compromises were necessary to maintain a concise presentation of results. For yields of DM, interaction means were sorted by planting date within year and reanalyzed as a split-plot design with cultivars as whole plots and harvest dates as subplots using PROC MIXED (SAS Institute, Cary, NC). For most of the nine combinations of year and planting date, the cultivar x harvest date interaction was significant; therefore, these interaction means are presented and discussed.[1]
• This count is regressed on country-year variables that control for a country’s changing export capabilities, including country dummy variables, time dummy variables, a vector of time-varying characteristics of country j, and the Post Reform Dummy. The country characteristics included in the vector Hjt are the same as those used in the specifications presented in Table 3. Because the data cover only an 11 year time frame, as explained below, country-specific time trends are not included. For data, we utilize the U.S. trade database created by Feenstra et al. (2001).[2]

As noted above, this strategy is all about the analytical procedures used to carry out the research. So, the previous Goal (Methods Goal 2: Describing the Study) is more about what happened during the data collection, while this goal explains to the reader what happened AFTER the data were collected.

The purpose of Goal 3, Analyzing the Data, is to persuade your reader about the quality of the data analysis and make a claim that the study’s procedures have led to valid and credible findings. As a reminder, this communicative goal allows the writer to explain how the data have been analyzed (without describing the results of that analysis).

## Strategies for Writing about Methods Goal 3: Analyzing the Data

• Preparing the data
• Describing the data analysis
• Establishing credibility
We’ll now discuss each of these and provide some examples from published research.

### Methods Goal 3 Strategy: Preparing the Data

Preparing the data describes what was done to the data and how the data were prepared for analysis. That is, this strategy is used to explain data selection, collection, and preparation (e.g., sampling, screening, cleaning, inclusion/exclusion, correction). You will also mention any tools that you used to accomplish these processes. Also, you need to explain data manipulation (e.g., transforming, coding, tabulating, estimating) along with any tools you used. Below are two examples of the utilization of this strategy in published research:

## Examples

• For all variables, least squares means were generated and, when significant (P < 0.05) F values were observed, least squares means were separated with pairwise t-test (PDIFF option).[3]
• Each electronic nose measurement was associated with the average (one representative value) of the eight human panelists assessments. The determination of the average pleasantness rating also involved trimming the data. The data were scanned for values that did not support the general consensus.[4]

Although the examples show data that were included in the research, you may also mention which data were excluded from the analysis. The following are some examples from the Academic Phrasebank website that you might consider using as sentence starters:

 Criteria for selecting the subjects were as follows: Publications were only included in the analysis if… The participants in this study were recruited from … To identify X, the following parameters were used: … The area of study was chosen for its relatively small … Primary inclusion criteria for the X participants were … Eligibility criteria required individuals to have received … Five individuals were excluded from the study on the basis of … A small sample was chosen because of the expected difficulty in obtaining … The subjects were selected on the basis of the degree of homogeneity of their … A comparison group of 12 male subjects without any history of X was drawn from a pool of …

### Methods Goal 3 Strategy: Describing the Data Analysis

Describing the data analysis provides a description of the actual analysis (with/without certain tools) in terms of how the data analysis was done and what procedures were used for analysis (e.g., statistical techniques, coding schemes, etc.).
Consider the following examples from published research:

## Examples

• For all variables, least squares means were generated and, when significant (P < 0.05) F values were observed, least squares means were separated with pairwise t-test (PDIFF option). [5]
• Our paired survey from buyers and suppliers significantly reduces the single-side, single-informant related common method variance bias. Before computing the average score from both sides, we also took steps recommended by Podsakoff and Organ (1986) and conducted a global factor analysis on items related to all predicting and criterion variables for each side. No single factor emerges from the analysis and no one factor accounts for most of the covariance for all predicting and criterion variables, confirming the absence of the common method bias. [6]

### Methods Goal 3 Strategy: Establishing Credibility

Establishing credibility means that you provide a rationale for the analysis and/or data processing and indicates statistical or other procedures employed to ensure credibility (e.g., reliability calculations). The implementation of this strategy gives recognition to existing and/or pre-existing limitations and/or explains or interprets certain observations or measurements.

Here are two examples taken from published research articles:

## Examples

• We only tentatively interpret the A mar data here, as its determination is complicated by uncertainties associated with the partitioning of marine and terrestrial deposits in the macrostratigraphy database (a subject of ongoing work).[7]
• Lastly, MM5 is intended to be a very conservative dissonance indicator, preventing potential misclassification of residents as mismatched as much as possible.[8]

## Key Takeaway

Overall, this goal provides information about how the researchers collected, manipulated, screened, cleaned, coded, and analyzed their data.

1. Coblentz, W. K., Bertram, M. G., & Martin, N. P. (2011). Planting date effects on fall forage production of oat cultivars in Wisconsin. Agronomy Journal103(1), 145-155.
2. Branstetter, L., Fisman, R., Foley, C. F., & Saggi, K. (2011). Does intellectual property rights reform spur industrial development?. Journal of International Economics83(1), 27-36.
3. Sawyer, J. T., Apple, J. K., Johnson, Z. B., Baublits, R. T., & Yancey, J. W. S. Color Stability of Dark-cutting Beef Enhanced with Lactic Acid. Arkansas, 87.
4. Williams, A. L., Heinemann, P. H., Wysocki, C. J., Beyer, D. M., & Graves, R. E. (2010). Prediction of hedonic tone using an electronic nose and artificial neural networks. Applied Engineering in Agriculture, 26(2), 343-350.
5. Sawyer, J. T., Apple, J. K., & Johnson, Z. B. (2007). The impact of acidic marination concentration and sodium chloride on sensory and ınstrumental color characteristics of dark-cutting beef. Arkansas Animal Science Department Report, 92-95.
6. Liu, Y., Luo, Y., & Liu, T. (2009). Governing buyer–supplier relationships through transactional and relational mechanisms: Evidence from China. Journal of Operations Management, 27(4), 294-309.
7. Meyers, S. R., & Peters, S. E. (2011). A 56 million year rhythm in North American sedimentation during the Phanerozoic. Earth and Planetary Science Letters, 303(3-4), 174-180.
8. Schwanen, T., & Mokhtarian, P. L. (2005). What affects commute mode choice: neighborhood physical structure or preferences toward neighborhoods? Journal of Transport Geography, 13(1), 83-99.