NASS Review of OMB 0596-0129

NASS Review of OMB 0596-0129

1. Sample selection

Forest Selection appears to follow a “management request” method. The target population is then all actual visitors to the particular forest and the statistical results then apply to the selected forest only and not to the general “Users of urban proximate National Forests” population. This study will miss the significant feature of potential visitors and their perceptions and reasons for not visiting the particular Forest. The National Survey methodology, referred to in the docket, would more likely address differences in perceptions between potential and actual visitors.

2. Site selection

The site selection portion of the docket is unclear to me. My general impression is that sites randomly selected with equal probabilities will not yield satisfactory results with respect to statistical validity or managerial decision making. For example, the docket refers to a Forest with 12 sites, each site with equal probability of selection, and then the site is replaced in the sample list for another possible selection. It is possible, though not likely, to exhaust the sample selection process and have selected only one site! The results obtained would apply to the particular site only and not be representative of the cumulative sites actual visitors may access in the given forest. Management decisions based on this sample selection would not be adequate for the particular selected forest.

3. Participant selection

Participant selection within a site has a large exposure to interviewer bias. Systematic selection is good when the sampled population is arranged in order and fixed, not spread out over a site. Also, entry and exit from the site cannot be controlled.

4. Data Collection

From my experience, a data collection cost of $48,000 for 600 responses for an instrument of this type is excessive.

5. Survey Instrument

Form A, question 5 and 6: Is there a statistical difference, in past studies, between what a responded believes is important for management and what the responded believes there is a need for more or less of in the forest?

All Forms, income question: Question wording….recommend, “What is your gross annual household income?”

All Forms, education question: Is there a statistical difference, in past studies, that justifies the Elementary, Middle, and High School subcategories?

All Forms, language question: Is this question attempting to measure information sources as they related to park usage and perceptions? If so, what about offering the sources for the respondent to select.

All Forms, birth place question: Does this really yield useful results? How about a collection of some residential variable such as zip code, county, city, etc.?

Form B: Questions with respect to mitigation of inappropriate uses and depreciative behavior. Shouldn’t these issues be based on enforcement economics and not user perception? Destruction of public property should be mitigated because it’s the right thing to do and not on visitor perception. Same thing for best management practices such as prescribed fire burns, safety, etc.

6. Docket Issues

Table 4 uses median wage estimates for U.S. Would a local statistic for median wages be more appropriate? What about calculating some value based on previous surveys with positive information in the income question?

Table 5 has no footnote for the asterisk in “Hourly Rate” column.

7. Recommendations

Sample selection based on weighted date, hour, and park entry grid. Ensures every element has positive probability of selection. Instrument supplied upon entry or exit to be mail returned using Business Reply. At twenty percent response, 3,000 mail returns are needed for a data collection cost of about $600.