September 28, 2018
By Lisa Petrison, Ph.D.
The Paradigm Change visitor map project provides information about locations in the U.S., based on the individuals who have visited the Paradigm Change sites (including the Paradigm Change website, the Paradigm Change blog and the Living Clean in a Dirty World blog).
This page summarizes information about the project and provides some guidelines on how the information possibly may be used for mold avoidance purposes.
The following map includes data about all U.S. visitors to the Paradigm Change sites from March 1, 2017 through September 15, 2018. A total of 202,082 different visitors are summarized with this map.
The color code for the 2018 map is as follows (and also is listed at the bottom of the map).
Black – Adjusted Index 750+
Fuchsia – Adjusted Index 500-749
Orange – Adjusted Index 300-499
Yellow – Adjusted Index 200-299
Green – Adjusted Index 100-199
White – Adjusted Index Below 100
This chart provides a variety of information about the largest 316 U.S. cities (those with 100,000 residents or more). It incorporates information about all U.S. visitors to the Paradigm Change sites between March 2017 and September 2018.
This chart provides information about visitors to the Paradigm Change sites by U.S. state. It incorporates information about all U.S. visitors to the sites between March 2017 and September 2018.
The following map includes data about all U.S. visitors to the Paradigm Change sites from September 2013 through June 2017. A total of 191,587 different visitors are summarized with this map.
The following map includes data about all U.S. visitors to the Paradigm Change sites over the 12-month period from July 2016 through June 2017. A total of 111,774 different visitors are summarized with this map.
The color code for the 2017 maps is as follows (and also is listed at the bottom of the maps).
Black – Adjusted Index 750+
Fuchsia – Adjusted Index 450-749
Orange – Adjusted Index 300-449
Yellow – Adjusted Index 150-299
Green – Adjusted Index Below 150
Light Green – Adjusted Index Below 150; 0-3 visitors
Light Grey – Adjusted Index 150+; 1-3 visitors
This chart provides a variety of information about the largest 316 U.S. cities (those with 100,000 residents or more). It incorporates information about all U.S. visitors to the Paradigm Change sites between September 2013 through June 2017.
This chart provides information about visitors to the Paradigm Change sites by U.S. state. It incorporates information about all U.S. visitors to the sites between September 2013 and June 2017.
This chart provides information about visitors to the Paradigm Change sites by Canadian province. It incorporates information about all Canadian visitors to the sites between September 2013 and June 2017.
A previous report about the Paradigm Change visitor map project is at the following link. It incorporates information about all visitors to the Paradigm Change sites between September 2013 and June 2016.
2017 Project Overview
The Paradigm Change visitor maps provide information about locations in the U.S., based on the individuals who have visited the Paradigm Change sites.
Data was obtained via Google Analytics for the periods of time stated in the descriptions above.
Only the “New Users” information provided by Google Analytics is included in the analysis. For this category of data, it is stated that no matter how many times a particular computing device has visited the sites, it is only recorded once in the data.
According to the Google Analytics report, a total of 191,587 computing devices from the U.S. and 10,808 computing devices from Canada visited the Paradigm Change sites on one or more occasions during the period of time being investigated.
The maps present index scores for each town in the U.S. These index scores are controlled for population, as is described below (in the “Index Score Calculations” section).
Information for each town is presented on the map in the following format:
NV Incline Village 1259 (67) – Adjusted 1259
The number following the information about state and town is the index score (in this case, 1259).
The index score provides a comparison between how likely people in the town were to have visited the sites vs. how likely people in the entire U.S. were to have visited the sites.
An index score of 100 suggests that the town had the same prevalence with regard to people visiting the site as was reported in the U.S. as a whole.
An index score of 200 suggests that the town had twice as high of a prevalence as was reported for the U.S. as a whole.
An index score of 1000 suggests that the town had ten times as high of a prevalence as was reported for the U.S. as a whole.
An index score of 50 suggests that the town had half as high of a prevalence as was reported for the U.S. as a whole.
For Incline Village, the index score of 1259 suggests that people in that town were a little more than 12x as likely to have visited the sites as were people in the U.S. as a whole.
The number in parentheses is the absolute number of visitors reported by Google Analytics to have visited the Paradigm Change site during the time period in question (in this case, 67 computing devices).
The adjusted index score is the number to the far right.
It re-calculates the index score looking only at the likelihood that individuals living in the town who are fluent in English visited the site, compared to the likelihood that all individuals in the U.S. who are fluent in English visited the site.
The 2017 maps use color-coding to identify the towns by their adjusted index scores.
Index Score Calculations
Information provided by Google Analytics was converted into index scores using the following calculations.
1. The population of each town in the US was determined based on 2010 census data.
2. The prevalence of visitors to the Paradigm Change site for each town was calculated.
This was done by taking the raw number of visitors from the town for the time period being considered (provided by Google Analytics) and dividing that by the population number for the town.
As an example, Dripping Springs, TX, had a total 2010 population of 1,788.
Google Analytics reported that 25 different computing devices (used here as a surrogate for number of people) in Dripping Springs visited the Paradigm Change sites at least once during the period from September 2013 – June 2017.
Therefore, 25 divided by 1788 is equal to .0139821 – an approximation of the percentage of people in Dripping Springs who visited the Paradigm Change sites during that period of time.
3. The prevalence of visitors to the Paradigm Change sites for the U.S. as a whole was calculated.
This was done by taking the total number of U.S. visitors to the Paradigm Change sites during the time period in question (191,587) and dividing it by the 2010 population of the U.S. (316,100,000).
The result was .000606096 – that is, an estimation of the percentage of people in the U.S. who had visited the Paradigm Change sites at least once during the designated time period.
4. The prevalence number for each town or state was divided by the prevalence number for the U.S. as a whole, and then multiplied by 100, to arrive at an index number.
For instance, for Dripping Springs, .0139821 was divided by .000606096, and then multiplied by 100, to arrive at an index score of 2307.
This index score suggests that people in Dripping Springs were about 23x as likely to visit the Paradigm Change sites at least once during that time period as were people in the U.S. as a whole.
Adjusted Index Score Calculations
The Adjusted Index score looks only at individuals in the town or state who speak English in the home.
It looks at how likely they were to visit the Paradigm Change sites, compared to how likely people in the U.S. as a whole who speak English in the home were to visit the Paradigm Change sites.
Following is the process by which the adjusted index scores were calculated for the 2017 map.
1. Information from the 2010 U.S. Census was used to determine the percentage of people speaking English in the home for each state.
This information was obtained from Table 4 of the following document.
In some cases, the document gave the percentage of people speaking English in the home for particular towns (Table 5). That data was used for this project when available; otherwise, the data provided for the whole state was used.
2. The document lists the percentage of people who speak a language other than English in the home. That percentage was subtracted from 100 to provide the percentage of people that do speak English in the home.
For instance, for Miami, 51.3% of the population does not speak English in the home, and thus 48.7% of the population does speak English in the home.
For the U.S. as a whole, about 20% of the population does not speak English in the home, and thus about 80% does speak English in the home.
3. To calculate the adjusted index, the original index was first divided by the proportion of the population speaking English in the home.
For instance, for Miami, the index score of 355 was divided by the percentage speaking English in the home for Miami (.487) to arrive at a preliminary adjusted score of 730.
4. The preliminary adjusted index score calculated in this way for the U.S. as a whole would be 125 (100 divided by .80).
5. The preliminary adjusted index for each town was divided by 125, and then multiplied by 100, to calculate the final Adjusted Index.
This number looks only at those people speaking English in the home, and compares the percentage of such people in the town in question to the percentage of such people in the U.S. as a whole.
For instance, for Miami, the preliminary adjusted score of 730 was divided by 125, and then multiplied by 100, to arrive at a final Adjusted Index of 584.
6. The final Adjusted Index score for each town was plotted on the map according to the color legend listed above.
U.S. cities with more than 100,000 residents are listed in the order of their adjusted index numbers on the following table.
As discussed in the “Interpreting the Data” section below below, it may be the case that on average, people who are less affluent will be less likely to make their way to the Paradigm Change sites than people who are more affluent. Therefore, towns with a median yearly household income below $60,000 are marked on the table as follows:
**** Median household income below $30,000.
*** Median household income below $40,000.
** Median household income below $50,000.
* Median household income below $60,000.
The right side of the table provides information about each town obtained from the Paradigm Change Locations Ratings project.
This is a Yelp-style rating project, in which individuals who already are pursuing mold avoidance rate locations where they have lived or visited in terms of the apparent effects that these places have had on their health. It uses a classic 5-star rating system (5 = Excellent, 4 = Good, 3 = Fair, 2 = Poor, 1 = Awful).
For instance, the table shows that San Francisco has been rated by 13 individuals and has an average rating of 2.0.
Interpreting the Data
Obviously, some towns have much higher adjusted index scores than others. Following is a list of possible reasons why that might be.
1. Random Error
One reason that some towns might register as having a high prevalence is simply because of random chance. It could be coincidence that an unexpectedly large number of people from a particular town found their way to the Paradigm Change site during the time period in question, rather than its being indicative of anything in particular about the town.
The large data set (almost 200,000 U.S. visitors) decreases the likelihood that random error is responsible for most the results. Still, considering the large number of towns being considered, the likelihood that at least a few of the effects observed are due to solely random error seems fairly high.
One thing to keep in mind when considering whether random error may be driving a particular result is the overall pattern of the data.
For instance, if a cluster of high-prevalence towns occur in close proximity to one another, then that may more indicative that random error is not to blame than if a high-prevalence point is found in isolation.
Also, higher-population towns with many site visitors may be more stable from a statistical point of view than lower-population towns with only a few site visitors.
Another topic to consider is whether the observed effects are consistent with regard to what experienced mold avoiders have said about an area or about what we know about it in general.
The more consistent that points on the map are with other information that we have about particular locations, the more likely those points would seem to be registering actual effects rather than random error effects.
2. Data Recording Inaccuracies
As noted above, Google Analytics is not 100% accurate in terms of its reporting of locational data.
In most cases, the errors seem to result in individuals being misattributed as being present a nearby town.
It therefore may be that looking at patterns of data for larger geographic areas may have the potential of being more accurate than would be using the data to pick out the best or worst towns within a particular area.
Interpretation problems may be especially likely to occur when a single town has a much higher prevalence level than other towns surrounding it, since it may be unclear whether this is due to 1) random error, 2) many users in the area being inaccurately registered as being from that town (e.g. because a large Internet service provider is located there), or 3) people from the town actually being especially inclined to visit the site (e.g. because they are especially likely to be sick with mold illness or related conditions).
3. Demographics and Psychographics
Certain kinds of people are especially likely to use the Internet to seek out health information on mold illness or other similar topics.
Therefore, the types of people that live in a particular town may make a difference in terms of their likelihood to visit the Paradigm Change site.
About 80% of Americans have Internet access at home. Poorer people disproportionately do not have access, and so towns with large numbers of disadvantaged people may be expected to come up lower on the map.
On the other hand, more educated and sophisticated individuals may be especially likely to use their computers to seek out health information and therefore to have been more likely to visit the Paradigm Change site than were average Americans.
In addition, since mold illness is not yet a wholly mainstream topic, it is possible that those who are more attuned to health trends or general trends may have been more likely to have visited the site during the time period in question.
In general, then, towns that have many people who are upscale, highly educated and/or open to trendy information conceivably may be expected to have had more than their fair share of residents visiting the Paradigm Change site and therefore to be represented with higher index numbers.
Towns that have many people who are disadvantaged economically, less educated and/or resistant to newer thinking may be expected to have had fewer than their fair share of residents visiting the Paradigm Change site and therefore to be represented with lower index numbers.
4. Illness Mecca Spots
In some cases, people with chronic illness may move to or visit particular locations because they are attracted to benefits that may be obtained there.
For instance, mold illness or MCS sufferers have been known to move to particular spots because they believe that living in those locations will be good for their health.
Other illness sufferers have moved to particular locations because healthcare practitioners who they want to see are located nearby.
Illness sufferers who have moved to these desirable locations may use the Paradigm Change site themselves and also may inform their new neighbors about the site, thus increasing site traffic from those towns.
Any of these factors may result in locations felt to be desirable by at least some illness sufferers coming up as higher-prevalence spots on the visitor map.
5. Environmental Toxicity
Environmental toxicity in certain locations has the potential of increasing the extent to which people in particular towns suffer from illness symptoms and thus seek out information from the Paradigm Change site.
For instance, some areas may have a larger-than-average number of particularly moldy buildings, as a result of factors such as shoddy construction, poor maintenance, or previous flooding.
Outdoor mold toxins or cyanobacteria toxins may be a factor in certain locations.
Other types of environmental toxicity (such as air pollution, agricultural chemicals, toxins related to oil and gas drilling, radiation, EMF’s or water contamination) may be a factor in some locations.
In some locations, combinations of these different types of environmental toxicity may work together to create a particularly problematic situation.
Towns rated as high-prevalence on the map should not necessarily be assumed to have environmental problems that are making people sicker and thus driving people to visit the Paradigm Change site. Other factors might be responsible.
However, insofar as points on the map seem to be a part of a general pattern, and insofar as other explanations do not seem to be wholly responsible for the patterns observed, then the idea that environmental issues may be playing a role may be worth considering.
6. Other Illness Drivers
In addition to environmental pollutants, other illness-promoting factors have the potential of being associated with particular locations.
For instance, environmental pathogens such as tickborne illness (e.g. Lyme disease or babesia) or fungal infections (such as Valley Fever or histoplasmosis) may play a role in causing people in some locations to be more likely to be sick with chronic multi-symptom illness and therefore more likely to visit the Paradigm Change sites.
In some cases, such as with the Amish, people who share particular genotypes may be more likely to live in particular locations. Insofar as these genotypes make people more susceptible to this type of chronic illness, this could result in those locations having particularly high index scores on the data map.
People in particular locations also may be more likely to share other risk factors that make them more likely to suffer from chronic illness and therefore to be more inclined to visit the Paradigm Change sites.
Some of these other factors could include consuming illness-promoting diets; getting less than an optimal amount of exercise; being overweight; getting less-than-optimal healthcare; having been previously treated with particular vaccines or pharmaceuticals; or engaging in particular lifestyle behaviors (such as smoking or recreational drug use).
One important thing to note about the Paradigm Change map is that despite the fact that it is controlled for population size, the correlation between high-prevalence spots and areas of high population density is extremely high.
This is consistent with the reports of many mold avoiders that they do better in more pristine areas away from civilization (sometimes termed “civilidevastation”), apparently as a result of the lesser amount of environmental toxicity in those less populated places.
The Paradigm Change visitor map also looks similar to many other maps that have been developed to depict areas of high environmental toxicity. Some of these other maps focus on conventional air pollution, mercury pollution, light pollution, noise pollution, cell phone radiation pollution, and (especially) hazardous chemical spills.
The total lack of high-prevalence locations in large sections of less developed states in the Western half of the US – and the systematic clustering of many points in most urban areas throughout the U.S. – also is worth noting.
Perhaps even more intriguing is the frequency with which points marked in black or fuchsia (Index 450+) cluster together in “hot spots” on the map.
These are extremely elevated prevalence numbers that do not seem to be necessarily associated with normal city pollution, humid areas, agricultural regions or other obvious factors associated with environmental toxicity issues.
Some big clusters of hot spots can be found in the countryside of Pennsylvania (just west of Philadelphia); in the semi-rural region just north of Atlanta; in much of southern Michigan (including the Ann Arbor area); and in parts of northern California.
Hot spots account for a high percentage of total visitors to the Paradigm Change site, thus suggesting the possibility that “ME/CFS” has remained in large part a cluster-driven disease.
Locations Ratings Project
As mentioned above (in the section on “Largest Cities”), a totally separate project presented from Paradigm Change is the Locations Ratings project.
This project asks those who have been pursuing mold avoidance to rate the apparent effects that locations where they have been have had on their health, using a 1-5 scale.
5 = Excellent
4 = Good
3 = Fair
2 = Poor
1 = Awful
Average ratings for towns with two or more raters are plotted on the Locations Ratings Map.
Data from the Locations Ratings project was correlated with data from the Paradigm Change Visitor project, to see how closely they were related to one another.
The number of people who have rated particular towns for the Locations Ratings project ranged from 0 raters to more than 11 raters.
Following are the correlations between the Locations Ratings project average rating and the Paradigm Change Visitor project adjusted index for cities with a population size of 100,000+, broken down by the number of raters.
2+ Raters (164 towns): r = -0.20
3+ Raters (112 towns): r = -0.20
4+ Raters (78 towns): r = -0.24
5+ Raters (59 towns): r = -0.26
6+ Raters (40 towns): r = -0.33
7+ Raters (33 towns): r = -0.38
8+ Raters (24 towns): r = -0.41
9+ Raters (19 towns): r = -0.51
10+ Raters (16 towns): r = -0.42
11+ Raters (12 towns): r = -0.59
12+ Raters (10 towns): r = -0.80
Following are the correlations between the Locations Ratings project average rating and the Paradigm Change Visitor project adjusted index for all towns, broken down by the number of raters.
2+ Raters (805 towns): r = -0.01
3+ Raters (384 towns): r = -0.08
4+ Raters (235 towns): r = -0.05
5+ Raters (144 towns): r = -0.02
6+ Raters (86 towns): r = +0.02
7+ Raters (53 towns): r = +0.07
8+ Raters (35 towns): r = -0.04
9+ Raters (23 towns): r = -0.16
10+ Raters (17 towns): r = +0.05
Interpretation of correlation effect sizes typically are categorized as follows:
Small: r = 0.10
Medium: r = 0.30
Large: r = 0.50
That being the case, it seems to be that the correlation sizes between the Locations Ratings numbers and the Paradigm Change Visitor Project numbers are moderately to strongly negatively correlated for larger cities (that is, that having a low rating is correlated with having a high index number).
For all towns, however, the correlation between the Locations Ratings numbers and the Paradigm Change Visitor Project numbers are only very weakly correlated.
Although it is difficult to know exactly why this would be, here are some of my own speculations on the topic.
First, even if we assume that all the variance in the Paradigm Change Visitors Project is due to how problematic the location is from a health perspective, it is still measuring a different construct than the Locations Ratings project.
The Paradigm Change Visitors Project seems that it may be driven at least in part by the extent to which people living in particular locations have become ill as a result of living in the location. It therefore potentially may be identifying toxins that contribute to the likelihood that people will become sick with the illness (e.g. that serve as risk factors for acquiring the illness).
The Locations Ratings project, on the other hand, has individuals who already have acquired mold-related illness and who already have started mold avoidance rate locations in terms of health effects – that is, on how much they trigger symptoms.
In terms of the Paradigm Change Visitors Project, the locations that are especially high prevalence (e.g. Adjusted Index 450 or higher) mostly seem to be ones that have been reported to be or that I have found to be especially problematic with regard particularly problematic environmental toxins (sometimes described as “super toxins”).
In some cases, these are areas that have been hit with substantial flooding and therefore have many moldy buildings.
However, in most cases, these locations have been reported as having substantial amounts of outdoor toxins (such as the sewer-based toxins referred to as “Mystery Toxin;” the toxins thought to be associated with fire retardants or fracking; the highly cross-contaminating toxins sometimes referred to as “Hell Toxin;” and particularly toxic cyanobacteria).
The Locations Ratings data and the Paradigm Change Visitor data correlate (inversely) fairly well when it comes to U.S. cities with more than 100,000 people. My feeling about this is that many cities tend to have a fairly similar mix of toxicity with regard to regular air pollution, meaning that the main thing that may distinguish whether some feel better than others to sensitized people may be the extent to which particularly problematic “super toxins” are present as part of the mix.
On the other hand, smaller towns often have an idiosyncratic mixture of toxicity, such as agricultural toxins or many other toxic issues. People who are environmentally sensitive may be triggered by this type of toxicity, even if it is not the kind of toxicity that would cause them to become sick with mold-related illness.
Personally, I found after a few years of avoidance that I was mostly reacting only to “super toxins,” and so my own ratings of various locations have been based pretty much solely on these toxins rather than on other more random factors.
When I correlated my own ratings with the adjusted index numbers from the Paradigm Change Visitor Project, it turns out that the (inverse) correlations were pretty high across the board.
For cities with more than 100,000 residents (55 ratings), the correlation between my ratings and the Paradigm Change Visitor Project adjusted index numbers was -0.50.
For all towns that I rated (191 ratings), the correlation with the Paradigm Change Visitor Project adjusted index numbers was -0.32.
I therefore think that it may be that my overall impression that the Paradigm Change Visitor Map is particularly good at identifying locations that have “super toxin” problems may be a theory with some validity to it.
The main problem that I see with using the maps for the purpose of identifying locations with super toxin issues is that there are a few spots with high prevalence levels that I do not believe to be especially problematic.
In some cases, these appear to be “mecca locations” – that is, smaller towns in good areas that many people with mold or chemical sensitivities have purposely moved to.
Several towns in Arizona (including Sedona, Cottonwood and Vail), New Mexico (Taos) and Colorado (Pagosa Springs) seem to fall into this category, for instance.
Also, Show Low, AZ, is a small remote town with very good air quality. I therefore wonder if its high adjusted index (826) may be driven by the fact that a community of individuals with environmental illness lives in nearby Snowflake, AZ. Since Google Analytics does not report any visitors at all as being located in Snowflake, perhaps it is the case that all site visits from this EI community are being misattributed to Show Low.
Another town with a very high prevalence level (adjusted index 3357) is Geneseo, IL. I have not received any reports that this is an unusually problematic town from an environmental perspective, and therefore am guessing that the large numbers of visitors that are attributed to the town are due to the fact that there is a large internet service provider there.
I therefore would like to suggest that it is not necessarily going to be a good idea to assume that a high prevalence level means that a location definitely is going to be problematic. There are some particularly prominent apparent exceptions.
In most cases though, I do think that a very high prevalence level (indicated by a black or fuchsia marker on the map) does signify that a town probably has a particularly problematic toxicity issue associated with it, and that this is especially likely to be the case when other nearby towns also have elevated prevalence levels.
Mold Avoiders Polls
Two polls in the Mold Avoiders group asked members to choose the worst locations for their health out of all the places that they had been. The results of the polls are summarized in these two Living Clean blog articles.
Data from the Paradigm Change Visitor Project are summarized along with the Worst Locations Poll data in this chart:
Part A of the chart lists the top 28 U.S. cities (population 100,000 or more), according to Adjusted Index Numbers.
The first column of numbers lists the Adjusted Index from the Paradigm Change Visitor Project.
The next two columns list the total number of Mold Avoiders members who stated that the town was one of the worst locations that they had been in either the 2018 or the 2016 poll.
The column of numbers to the far right lists the total number of individuals who stated in the polls that the town was among the worst ones they had ever visited or lived, combining the 2016 and 2018 data.
Part B of the chart lists additional U.S. cities (population 100,000 or more) for which at least two individuals total stated in the polls that it was one of the worst locations that they had ever visited or lived.
Part C of the chart lists smaller towns (population less than 100,000) for which at least two individuals stated that it was one of the worst locations that they had ever visited or lived.
In addition, an analysis was done for the 316 largest U.S. cities (population 100,000 or more), comparing the total number of people who stated that each town was the worst location they had ever been with the Adjusted Index Number.
The correlation result was r = 0.61, which is considered to be a very large effect size.
It therefore seems that at least for the largest cities, the information provided on the Paradigm Change Visitor Project appears to be fairly accurate at identifying locations that Mold Avoiders members feel have had a particularly negative effect on their health.
This is consistent with how I have personally been using the Paradigm Change Visitor Project – to predict which locations have the potential of being particularly problematic, so that I can either avoid these places or exercise particular caution when visiting them.
Thus far, I feel that the map has been helpful to me for this purpose.
Unfortunately, Google Analytics lists data by town only for the U.S. rather than for other countries.
For Canada, information about site visitors is provided for each province. It is summarized in this table.
To prepare the information for the chart, the prevalence of individuals from each province who visited the Paradigm Change sites was compared to the prevalence of individuals from the U.S. as a whole who visited the Paradigm Change sites.
Only regular Index Numbers (rather than Adjusted Index Numbers) are listed on this chart.
As a result, it likely is the case that the very low index number for Quebec (a Canadian province where French is used as the primary language) is due to the fact that many people in Quebec may not feel comfortable navigating the Paradigm Change site (which is totally in English), rather than to the province being especially good in terms of environmental issues.
Following are links to the Paradigm Change sites.
Discussions of the Paradigm Change Visitor Project take place on the Mold Avoiders group on Facebook:
Please direct comments and questions about this project to Lisa Petrison at the following address:
info at paradigmchange dot me
Links on this page are in orange (no underlining).