Prior to conducting analysis, the data should be examined for anomalies. A good first step is to review the histograms for each column. Select Analyze > Distribution and enter all of the columns into the Y Columns field.

Figure 2.1 Histogram for Primary Role shows the histogram for the column Primary Role, which reveals that three Licensed Practical Nurses (LPN) responded to the survey.

Only registered nurses were eligible to take the survey, therefore the LPN responses should not be included in the analysis. Associate nurse managers, nurse educators/quality specialists, and nurse managers are required to be registered nurses. The rows containing the LPN responses can be found using Row > Data Filter, as shown in Figure 2.2 Data Filter to Select LPN Responses.

The data table will have three rows highlighted. Right-click one of the highlighted rows and select the Exclude/Unexclude toggle. Excluding a row will omit that observation from calculations, but it will be shown on JMP graphs. Observations can be omitted from graphs using the Hide/Unhide toggle. An observation can be both excluded and hidden with the Exclude and Hide toggle. The presence of three ineligible survey responses warrants investigation into the survey design and distribution process.

Figure 2.3 Histogram for Primary Role (2) shows the histogram for Primary Role for the eligible survey respondents.

It is important to give the response rate when presenting survey results. The Distribution output for the consent question, “Do you agree to participate in the survey?” is shown in Figure 2.4 Distribution for Survey Consent.

There are 262 affirmative responses and 22 missing values. The missing values exist because of the way in which the survey tool operates. As soon as the application is opened a record is created. If a respondent does not consent to participate, the survey application closes without recording a value in the “Consent” column. When importing data into JMP that is created by other software, it is important to understand how the data records are created. For this survey, 22 respondents who opened the survey tool did not consent.

Taking into account the three ineligible LPN responses and the 22 nurses who did not consent to the survey there are a total of 259 valid responses and the correct response rate is 259/854 = 30%. It will be easier to find the non-response rates for the various survey questions to create a new data table that contains only the valid survey responses. This can be done by selecting the rows corresponding to the valid responses and creating a new table using Tables > Subset.