reducing threats to internal and external validity. They
also take steps to reduce error when deciding whether a
research hypothesis is supported or not supported.
When making decisions about hypotheses, research-
ers keep several principles in mind (Hayes, 1994). First,
hypotheses are claims about the world. Second, decisions
are made about null hypotheses, not research hypotheses.
The purpose of inferential statistical tests is to determine
whether the null hypothesis should be accepted or rejected. Third, it is impor-
tant to remember that no matter how well errors are reduced or how powerful
the findings are, nothing is ever proven. The most that can be claimed is that
a research hypothesis is either supported or unsupported. Last, it is wise to
understand that an empirical view is adopted and an assumption is made that
there is a single reality in the physical world and that science can be used to
discover the truth about reality.
For example, suppose a researcher is testing a new way to assess risk for
skin breakdown over the sacrum. The researcher hypothesizes that prior to the
presentation of redness, skin temperature will be increased. From an empirical
viewpoint, it is assumed that skin temperature, redness, and sacrum are real
entities and that in the real world one of four situations is true: prior to the
appearance of redness over the sacrum, the skin temperature will be increased,
unchanged, decreased, or varied. Now suppose that in the real world, the
unknown truth is that skin temperature does increase before redness appears
over the sacrum. After collecting and analyzing data, statistically significant
results were obtained. The researcher decides to reject the null hypothesis,
that there will be no difference in skin temperature before the appearance of
redness, and the research hypothesis is supported. In this instance, unknown
to the researcher, an error was not made. The claim of the original hypothesis,
that skin temperature would be increased prior to the appearance of sacral
redness, matches what actually happens in the world. The researcher rejected
the null hypothesis, and it was indeed false. If nurses accumulated additional
supporting evidence, findings would make their way into practice, and patients
at risk for skin breakdown could be identified earlier.
Errors are avoided when researchers accept the null hypothesis when it is true.
Consider this variation of the preceding example. The researcher hypothesizes
that prior to the presentation of redness over the sacrum, skin temperature will
be increased; however, in the real world there is no change in skin temperature
over the sacrum prior to the appearance of redness. After collecting and analyz-
ing data, the researcher finds that calculations from inferential statistical tests
are nonsignificant and therefore the null hypothesis is supported. Although
disappointed, the researcher accepts the null hypothesis and unwittingly avoidsFYI
Researchers attempt to reduce error so
that nurses can have confidence in findings.
Methods for reducing error include selecting
designs that fit with research questions, con-
trolling the independent variable, carefully
measuring variables, and reducing threats
to internal and external validity.356 CHAPTER 13 What Do the Quantitative Data Mean?