Ecological Archives A012-006-A1

Christy A. Brigham, Alison G. Power, and Alison Hunter. 2002. Evaluating the internal consistency of recovery plans for federally endangered species. Ecological Applications 12:648-654.

Appendix A. Consistency measures for questions 1-9.

The mean percentage consistency for each question was computed by dividing the number of consistent responses by the total number of responses for a given question. 

1) There were 59 different threat categories that could be identified as being present in the listing document (Questions 173-209 column L in the questionnaire).  These were compared for consistency with column M for the same questions.  Thus there were 59 different opportunities for consistency for each plan.  If a threat was mentioned in the listing but not in the recovery plan, it was considered an inconsistent response.

2) There were 59 different categories of threat (Questions 173-209) that could be cited within the plan as a threat.  These 59 categories could also be assigned recovery tasks (Questions 235-293).  A "yes" for a cell in column M that was paired with a "yes" in the same cell in column X was considered a consistent response. 

3) A "yes" in column M with no response in column X was an inconsistent response.  To match up threats with monitoring tasks, some questions from the questionnaire were lumped.  Lumping was done to pair categories of threat with the appropriate monitoring category (for example, if exotic species were identified as a threat in Questions 209-215, this was paired with a task to monitor exotic species, Question 364).  The following threat categories (given here as the question number under which they appear in the questionnaire) were paired with the following task categories.  A response greater than zero in the task category was considered a consistent response.

Column M
Column OO
209-215 364
216, 209 362
217, 210 361
211-213, 218, 220 363
219 360
222-226 368
173-189, 193-207 366, 368

4) Major threats were identified from column P (any response that received a "2").  These tasks were examined to see whether they received a high implementation priority.  A prioritization of a "1" or "2" was considered a consistent response.  A priority "3" assignment was considered an inconsistent response.  The same pairings shown in Question 5 were used for Question 6.

5) There were 38 different categories of information that could have been missing for a species (Questions 89-126 in the questionnaire).  For these questions column A was compared to column F.  Column A asks whether information was provided for a given topic (e.g., habitat requirements) and question F asks whether more information in this area was requested.  Only questions with a "no" in column A (i.e.,  "no information was provided in this area") were considered.  A "no" in column A paired with a "yes" in column F was considered a consistent response.  The total number of columns with a consistent response was divided by the total number of cells with a "no" in column A to give a total percentage consistent for this question.  We did not include any plans with a "yes" in column A, thus we may be slightly underestimating the consistency for this question since some plans may contain information on a topic but still want more information.  These plans would be consistent but were not sampled.

6) There were 38 different categories of information that could have been missing for a species (Questions 89-126 in the questionnaire) and 38 matching cells to ask whether this missing information had a task associated with it (Questions 130-167 in the questionnaire).  Agreement between cells F and G for a given question was considered a consistent response.  A "yes" in column F matched with "0" tasks in Question G was considered an inconsistent response. 

7) If a measure was identified as a recovery criteria as identified as a response of "1 or more" in column EEE, then does it have a monitoring task associated with it in column OO?  If there was a "1 or more" in column EEE it must be associated with a "1 or more" in column OO to be considered a consistent response.  A "zero" or "no" response in column OO was considered an inconsistent response.  The following questions were paired based on category of information (for example a recovery criteria involving total population size, Question 348, must involve a task in the question involving monitoring total population size, Question 405).

Column EEE
Column OO
405 348
408 349
406 350
409 351
418 353, 354

8) Column FFF identifies whether a specific recovery criterion requires qualitative or quantitative data.  This column was paired with a corresponding column (RR) that identifies whether monitoring will discover the type of data needed.  Monitoring that provided the type of data needed to assess recovery was considered consistent.  Monitoring that did not provide the information needed was inconsistent. Pairing of questions was the same as in Question 7. 

9) If column FF (management tasks) was greater than zero it must be matched with a non-zero response in column OO (monitoring tasks) to be considered a consistent response.  The following questions were paired based on subject.

Column FF
Column OO
314 361
315 363
317 362
319 368
319-331 366-370


[Back to A012-006]