Biostatistics For Dummies, 2nd Edition (Monika Wahi, John Pezzullo)

224 PART 5 Looking for Relationships with Correlation and Regression

» ^{Figure 18-6 shows most of the residuals lie within}¹⁰^{mmHg of the line, but}

several larger residuals appear to be where the SBP is around 115 mmHg.

We think this is a little suspicious. If you see a pattern like this, you should

examine your raw data to see whether there are unusual values associated

with these particular participants.»

» ^{If the residuals are normally distributed, then in the normal Q-Q chart in}

Figure 16-6, the points should lie close to the dotted diagonal line, and

shouldn’t display any overall curved shape. Our opinion is that the points

follow the dotted line pretty well, so we’re not concerned about lack of

normality in the residuals.

Making your way through

the regression table

When doing regression, it is common to focus on the table of regression coeffi-

cients. It’s likely where you look first when interpreting your results, and where

you concentrate most of your attention. When you run a straight-line regression,

statistical programs typically produce a table of regression coefficients that looks

much like the one in Figure 16-4 under the heading Coefficients.

For straight-line regression, the coefficients table has two rows that correspond

with the two parameters of the straight line:»

» ^{The intercept row:}^{This row is labeled}^(Intercept)^{in Figure 16-4, but can be}

labeled Intercept or Constant in other software.»

» ^{The slope row:}^{This row is usually labeled with the name of the independent}

variable in your data, so in Figure 16-4, it is named Wgt. It may be labeled Slope

in some programs.

The table has several different columns, depending on the software. In

Figure 16-4, the columns are Estimate, Std. Error, t value, and Pr (>|t|). How to inter-

pret the results from these columns is discussed in the next section.

The values of the coefficients (the

intercept and the slope)

The first column of the table of regression coefficients usually shows the values of

the slope and intercept of the fitted straight line (labeled Estimate in Figure 16-4).

Other column headings include Coefficient or the single letter B or C (in uppercase

or lowercase), depending on the software.